Intracoronary Doppler Assessment of Moderate Coronary Artery Disease
Comparison With 201Tl Imaging and Coronary Angiography
Background Coronary angiography may not reliably predict whether a stenosis causes exercise-induced ischemia. Intracoronary Doppler ultrasound may enhance diagnostic accuracy by providing a physiological assessment of stenosis severity. The goal of this study was to compare intracoronary Doppler ultrasound with both 201Tl imaging and coronary angiography.
Methods and Results Fifty-five patients with 67 stenotic coronary arteries underwent coronary angiography with intracoronary Doppler ultrasound and had exercise 201Tl testing within a 1-week period. Coronary flow reserve was measured, and analyses were performed by independent core laboratories. The mean stenosis was 59±12%; 51 of 67 stenoses were intermediate in severity (40% to 70%). A coronary flow reserve <1.7 predicted the presence of a stress 201Tl defect in 56 of 67 stenoses (agreement=84%; κ=0.67; 95% CI=0.48 to 0.86). In the patients who achieved 75% of their predicted maximum heart rate, the Doppler and 201Tl imaging data agreed in 46 of 52 stenoses (agreement=88%; κ=0.77; 95%CI=0.57 to 0.97). Scatter was evident when angiography was compared with coronary flow reserve (r=.43), and the angiogram did not reliably predict the results of the 201Tl stress test (κ=0.21; agreement=57% to 63%).
Conclusions Doppler-derived coronary flow reserve accurately predicts the presence of exercise-induced ischemia on stress 201Tl imaging, and coronary angiography does not reliably assess the physiological significance of an intermediate coronary stenosis.
More than 1 million coronary angiograms are performed in the United States annually.1 However, angiography provides only a two-dimensional silhouette of the coronary lumen; this limited anatomic assessment may not reliably predict whether a stenosis causes exercise-induced ischemia.2 3 4 5 6 7 The angiogram may be most limited in its assessment of lesions of intermediate severity, defined as a percent diameter stenosis between 40% and 70%. Although angiographic assessments of intermediate coronary lesions may lead to inappropriate decisions regarding revascularization, exercise tests documenting ischemia are generally not performed before coronary angioplasty.8 The magnitude of this problem may be quite large when one considers that nearly 90% of lesions undergoing angioplasty have a diameter stenosis <70%.9
Impaired coronary flow reserve is a hallmark of a physiologically significant coronary stenosis10 ; this parameter served as the “gold standard” during the development and validation of both coronary angiography and 201Tl perfusion imaging in animal models.11 12 13 14 15 Until recently, coronary flow reserve could not be easily measured in clinical practice. Coronary angiography has therefore remained the clinical “gold standard” for diagnosing coronary artery disease, despite its widely acknowledged limitations. However, poststenotic blood flow velocity can now be measured with a Doppler-tipped angioplasty guidewire at the time of angiography.16 By inducing maximal hyperemia with an appropriate arteriolar vasodilator, coronary flow reserve (the ratio of maximal hyperemic/baseline blood flow velocities) can now be determined. This physiological assessment of stenosis severity, which is now available in the catheterization laboratory, may add important complementary information to the coronary angiogram.
We measured regional coronary flow reserve and performed exercise 201Tl tests in patients undergoing coronary angiography. The goals of this study were twofold: (1) to determine whether coronary flow reserve, measured at the time of angiography, could predict the presence of ischemia on exercise 201Tl single-photon emission computed tomography (SPECT) imaging and (2) to critically evaluate the relationship between coronary angiography, regional blood flow velocity, and stress 201Tl imaging in patients undergoing evaluation of moderate coronary artery disease.
Protocol and Study Sites
After providing informed consent, all patients had a coronary angiogram and a stress 201Tl SPECT study within a 1-week period. Intracoronary Doppler was performed at the time of angiography. Hospitals and investigators were selected on the basis of their experience with intracoronary Doppler ultrasound and familiarity with clinical investigations in both interventional and nuclear cardiology. The protocol was approved by the institutional review committees at the 10 participating centers.
Patients referred for coronary angiography or coronary angioplasty were screened for eligibility. Patients were eligible for the present study if they had a single stenosis in the proximal or mid portion of one or two major coronary arteries and could participate in a Bruce treadmill protocol. Patients were ineligible for the study if any of the following exclusion criteria were present: left-main/triple-vessel disease, prior bypass surgery, significant valvular heart disease (>2+ regurgitation by ventriculography, aortic valve area <1.0 cm2, or mitral valve area <1.5 cm2), left ventricular hypertrophy by ECG (Estes’ criteria),17 an ejection fraction <45%, or inability to obtain informed consent. Vessels that were totally occluded or supplied an area of prior myocardial infarction (defined by Q waves on the ECG and/or regional akinesia) were also not evaluated for the present study.
All patients had poststenotic Doppler recordings in at least one stenotic coronary artery and in the proximal portion of at least one nonstenotic reference vessel. Blood flow velocity was measured 1 cm proximal and 1 cm distal to the target stenosis with a Doppler guidewire (FloWire, Cardiometrics). The location of the Doppler guidewire was documented by cineangiography.
The patient’s heart rate and blood pressure were recorded during each Doppler measurement, and the rate-pressure product was calculated. Hyperemia was induced with two separate intracoronary adenosine injections (12 μg in the right coronary artery and 18 μg in the left coronary artery), and coronary flow reserve was then calculated as the ratio of the average peak velocity during maximal hyperemia to the baseline average peak velocity. Because coronary blood flow is closely related to the rate-pressure product,18 the coronary flow reserve was then corrected by multiplying the baseline average peak velocity by the ratio of the mean rate-pressure product for all the stenoses divided by the rate-pressure product for the individual stenosis. The ratio of the baseline average peak velocity proximal and distal to the stenosis (proximal/distal velocity ratio) was also calculated. Volumetric blood flow was determined by multiplying the interpolated area of the arterial reference segment (derived from the angiogram) and the mean velocity (Average Peak Velocity/2)19 ; the coronary resistance index (in millimeters of mercury per minute per milliliter) was then calculated by dividing mean arterial pressure by coronary blood flow. The flow ratio index during hyperemia was calculated as the ratio of the poststenotic coronary flow reserve to the reference artery coronary flow reserve. In patients with single-vessel disease, the highest coronary flow reserve in the two reference vessels was used to calculate the flow ratio index.
Treadmill exercise. Patients underwent a symptom-limited exercise test using the standard (53 of 55 patients) or modified Bruce protocol (2 of 55 patients). Patients remained fasting for 4 to 6 hours before exercise testing. At peak exercise, 3 to 4.5 mCi 201Tl was injected intravenously.
Thallium imaging. Within 7 to 10 minutes after termination of exercise, a stress anterior planar image was acquired for 5 minutes to assess pulmonary 201Tl uptake. This was immediately followed by acquisition of a cardiac SPECT study. Delayed 201Tl SPECT images were acquired 3 to 4 hours later. Imaging at the different sites was performed with the use of a variety of single-, double-, and triple-headed SPECT cameras. SPECT imaging was performed with low-energy, high-resolution collimators over a 180° arc, 30 to 60 projections per study using a symmetrical 20% energy window centered over the 80-keV x-ray peak and a 10% window over the 167-keV γ-ray peak. Images were acquired into a 64×64-pixel matrix. The projection images were stored on magnetic media and transferred to an Odyssey workstation (Picker) after translation into Picker image format by use of Gammacon (MITA Inc) image-exchange software. Transverse images were reconstructed by use of filtered backprojection after low-pass Butterworth filter, order 5, frequency cutoff of 0.32 cycles/cm. The transverse images were then reformatted into short-axis, vertical long-axis, and horizontal long-axis slices for display on the Picker workstation.
Image analysis. Stress and delay images were assessed by the consensus of two nuclear cardiologists who were blinded to the Doppler and angiographic data. The perfusion in the distribution of each of the three coronary arteries was then rated as normal or abnormal and subsequently compared with Doppler coronary flow data and angiography.
Quantitative Coronary Angiography
The cineangiograms and distal segments of the angiographic catheters were forwarded to the Washington Hospital Center Angiographic Core Laboratory for analysis. Selected cine frames demonstrating the stenosis or proximal vessel in its two sharpest, least foreshortened projections were digitized by use of a cine video converter. The catheter outer diameter was directly measured; with the contrast-filled distal catheter used as the calibration standard, the minimal lumen diameter and reference diameter were then determined with the use of a validated commercial edge-detection algorithm (ARTREK).20 From these measurements, the percent diameter stenosis was determined.
The Doppler, nuclear, and angiographic studies were forwarded to independent core laboratories for blinded analysis. The Doppler laboratory at the University of Massachusetts analyzed original Doppler prints and hemodynamic data. The nuclear laboratory at the University of Massachusetts received both raw scan and treadmill data; the scans were reconstructed and interpreted as described above. The angiograms were sent to the core laboratory at the Washington Hospital Center for quantitative analysis.
Data Management and Statistical Analysis
All data were recorded by the research coordinators. The case report forms and core laboratory analyses were sent to the statistical group at the University of Massachusetts. Case report forms were audited at the coordinating center; data were verified by range and consistency checks and double data entry. A Data and Safety Review Committee composed of three outside experts in nuclear cardiology, interventional cardiology, and biostatistics, respectively, reviewed the study in an ongoing manner.
Descriptive analyses of the patient population were conducted using Proc Freq and Proc Univariate (SAS) for categorical and continuous variables, respectively. Continuous variables were expressed as mean±1 SD. Changes in hemodynamic characteristics and Doppler measurements were compared at baseline and during maximal hyperemia by use of a paired t test. Comparisons between Doppler measures in normal and abnormal 201Tl and angiographic distributions were made by use of a t test. The agreement between Doppler (dichotomized as normal and abnormal) and nuclear data and the agreement between nuclear data and angiographic data for individual vessels (after dichotomizing diameter stenosis at 50% and 70%) were assessed by use of the κ-statistic. κ is presented as the percentage of agreement with 95% confidence limits.21 The best “cut point” for comparing Doppler or angiographic data with 201Tl was identified as the value with the highest sum of both sensitivity and specificity. Scatterplots were used to assess the association between diameter stenosis and minimal luminal diameter with coronary flow reserve. The strength of the association was determined by use of the Pearson product moment correlation coefficient (Proc Corr, SAS). Because 55 patients contributed 67 stenoses to the analysis, we assessed the assumption of independence between stenoses from the same individual using regression modeling (Proc GLM, SAS). There was no violation of the assumption of independence.22
Patient and Angiographic Characteristics
Fifty-five patients with 67 stenotic coronary arteries were studied (Tables 1⇓ and 2⇓). Three quarters of the lesions studied (51 of 67) were characterized as stenoses of intermediate severity. Lesions were found in 36 left anterior descending, 15 left circumflex, and 16 right coronary arteries in the study group. Collaterals were not present.
Doppler and Hemodynamic Characteristics
The baseline and hyperemic heart rates, blood pressures, and Doppler-derived variables are shown in Tables 3⇓ and 4⇓. There was no significant change in the heart rate, blood pressure, or rate-pressure product after intracoronary adenosine administration; however, coronary resistance decreased during maximal hyperemia (P=.008). The mean poststenotic flow reserve for the 67 stenoses (1.9±0.8; range, 1.0 to 3.6) did not significantly change after correcting for the baseline rate-pressure product (1.9±0.8; range, 0.8 to 4.0).
Comparison of Doppler and 201Tl Imaging
The 67 stenotic arteries were associated with 35 normal regions and 32 defects on the stress 201Tl scans. The blinded 201Tl readings were compared with a variety of Doppler indices (Table 5⇓); however, the best agreement was found with the coronary flow reserve (κ=0.67; cut point=1.7). The results of the coronary flow reserve/201Tl comparisons are shown in Figs 1⇓ and 2⇓. Excluding patients who failed to achieve 75% of the predicted maximal heart rate during exercise improved the flow reserve/201Tl agreement but did not change the best cut point (κ=0.77; cut point=1.7). Correcting the coronary flow reserve for the rate-pressure product did not improve the level of agreement.
Comparison of Doppler and Angiography
The 67 stenoses were divided into two groups: <50% (n=14) and ≥50% (n=53). An unpaired t test demonstrated that the coronary flow reserve was higher distal to stenoses <50% (2.3±0.5 versus 1.8±0.8; P=.05). Least-squares regression analysis demonstrated a significant linear relationship between coronary flow reserve and both diameter stenosis and minimal luminal diameter (P<.001); however, both scattergrams revealed significant variability about the fitted lines (r=.43 and r=.41, respectively, for Fig 3A⇓ and 3B⇓). The relation was most variable in the intermediate range of stenosis severity; indeed, flow reserve values ranged between 1.0 and 3.6 for stenoses in the 50% to 60% range.
Comparison of Angiography and 201Tl Imaging
The 35 stenoses associated with normal 201Tl distributions were less severe than the stenoses in the 32 abnormal distributions (55±11% versus 64±12%; P=.001) The angiographic data were then dichotomized (Fig 4⇓). When ≥50% and ≥70% diameter stenosis were used as the definition of a significant stenosis, the agreements with the blinded 201Tl interpretations were 57% (κ=0.16) and 63% (κ=0.23), respectively. Restricting the analysis to patients achieving ≥75% of the predicted maximum heart rate did not significantly alter these results (58% agreement and κ=0.19 for 50%; 62% agreement and κ=0.19 for 70%). Although an evaluation of the sensitivity and specificity curves revealed a best angiographic cut point of 55% for this population, reanalysis of the data did not substantially improve the results (70% agreement and κ=0.40; 67% agreement and κ=0.33 in patients achieving ≥75% of the predicted maximum heart rate).
The primary findings of this study are that a Doppler-derived coronary flow reserve of <1.7 accurately predicts the presence of exercise-induced ischemia on stress 201Tl imaging, and coronary angiography does not reliably predict the physiological significance of an intermediate coronary stenosis as measured by either intracoronary Doppler ultrasound or stress 201Tl imaging.
Comparison of Doppler and 201Tl Imaging
The stress 201Tl scans were compared with a variety of intracoronary Doppler indices. However, of the various hyperemic indices studied, coronary flow reserve demonstrated the highest level of agreement. Because low-level exercise is a poor stimulus for hyperemic coronary flow, higher exercise heart rates improved the correlation between poststenotic flow reserve and exercise 201Tl testing.
It has been shown that a 50% reduction in regional flow during hyperemia is associated with the appearance of stress 201Tl defects in a canine model.15 However, the flow ratio index, a measure of regional flow heterogeneity during hyperemia, did not improve the ability of intracoronary Doppler to predict ischemia on a stress 201Tl scan. Coronary flow reserve measured in the proximal portion of a reference vessel may not be an accurate barometer of distal flow abnormalities, and changes in regional blood flow velocity may not always reflect volumetric flow heterogeneity during exercise.23 However, the finding that a 1.7:1 flow ratio index (corresponding to a 40% reduction of distal flow) is associated with stress 201Tl defects provides evidence that the conclusions previously drawn in animal models are applicable to clinical myocardial perfusion imaging.
Intracoronary adenosine administration resulted in a decrease in coronary resistance without a change in blood pressure or heart rate; however, the hyperemic resistance index and the stress 201Tl scans did not correlate. This finding may reflect the need to measure distal coronary artery pressure to calculate the true poststenotic coronary resistance or the limitations of calculating volumetric coronary flow by use of Doppler and angiographic parameters. There was also a poor correlation between the baseline proximal/distal velocity ratio and the stress 201Tl scan both in the left and right coronary arteries.24 One might expect changes in blood flow velocity during adenosine-induced hyperemia to better predict the presence of an exercise-induced 201Tl defect.
Comparison of Angiography and Doppler/201Tl Imaging
Although there was a significant linear relation between angiographic stenosis severity and intracoronary Doppler ultrasound, variability was noted when either diameter stenosis or minimum luminal diameter was directly plotted against coronary flow reserve. For example, a stenosis between 50% and 60% was associated with a poststenotic flow reserve ranging from as low as 1.0 to as high as 3.6.
The coronary angiogram also did not reliably predict the results of the stress 201Tl scan. Neither a 50% nor 70% diameter stenosis, the most widely used angiographic criteria for stenosis severity, reliably predicted the presence of a reversible defect on the stress 201Tl scan. The level of agreement was not improved by varying the angiographic cut point for stenosis severity or by excluding the patients who failed to achieve at least 75% of the predicted maximum heart rate during exercise. The relatively poor agreement between 201Tl imaging and angiography in this study reflects the inability of the angiogram to reliably predict the physiological significance of an intermediate coronary stenosis.
Relationship to Prior Investigations
Several small, single-center studies have reported a high correlation between coronary flow reserve and stress perfusion imaging. Early studies used older, less reliable methods for measuring coronary flow reserve.25 26 More recent investigations have produced conflicting results regarding both the angiography/coronary flow reserve relationship and the best coronary flow reserve cut point for determining whether a stenosis is flow limiting.27 28 29
The present study provides definitive clinical data from a large, multicenter experience that confirms the inability of angiography to predict the functional significance of intermediate coronary stenoses and establishes a coronary flow reserve of <1.7 as the best predictor of ischemia on a stress 201Tl scan. This protocol also used independent core laboratories, quantitative coronary angiography, and uniform stress imaging protocols. In addition, Doppler measurements in nondiseased reference vessels provided insights into the relationship between regional blood flow and perfusion imaging in patients with coronary artery disease.
This study primarily evaluated stenoses of intermediate severity; indeed, three quarters of the 67 stenoses studied were in the 40% to 70% range. One would expect a higher correlation between angiography and functional testing when the entire range of stenosis severity is evaluated. However, these data also demonstrate excellent agreement between coronary flow reserve and 201Tl imaging in the subset of stenoses that pose the greatest diagnostic challenge.
The majority of patients studied had single-vessel disease. Indeed, patients with triple-vessel disease were excluded from this study because they may demonstrate balanced hypoperfusion and false-negative stress 201Tl scans. Because 201Tl imaging served as the standard for lesion significance in this study, we chose a study population that maximized the reliability of this diagnostic modality.
Although commonly used for assessing epicardial stenoses, the coronary flow reserve is a hemodynamic index that theoretically reflects the sum of the stenosis, the microvasculature, and the distal myocardial perfusion bed. Likewise, 201Tl images may be abnormal in the presence of microvascular or myocardial disease. Despite these limitations, myocardial perfusion imaging is a time-honored test for the functional assessment of stenosis severity; the ability to acquire similar data at the time of angiography represents an important clinical advance.
Exercise 201Tl testing was done within 1 week of the intracoronary Doppler assessment; a comparison at the same point of time may have resulted in an even higher correlation. Furthermore, the local hyperemia produced by an intracoronary adenosine bolus may not be comparable to a lower level of global hyperemia produced by treadmill exercise. Despite these differences, Doppler/201Tl imaging concordance was seen in nearly 90% of stenoses when 75% of the predicted maximum heart rate was achieved.
This investigation was performed in patients with stable coronary artery disease who could perform a Bruce treadmill protocol. Because 201Tl uptake and flow reserve in the infarct zone may be abnormal, arteries supplying infarcted distributions were not evaluated. The exclusion of infarcted distributions allowed the relationship between stenosis severity and the three diagnostic modalities to be assessed in the absence of the confounding variable of viability. However, these results should be applied cautiously both to patients with unstable coronary syndromes and to infarct-related arteries in which an abnormal flow reserve may reflect an absence of viable myocardium.30
Transstenotic pressure gradients were not measured in this study, and coronary resistance was therefore estimated using mean aortic pressure. It has been suggested that transstenotic pressure measurements at baseline and during hyperemia may provide an assessment of stenosis severity that is independent of the status of the microvasculature.31 However, a recent study suggests that transstenotic pressure measurements are inferior to coronary flow reserve for predicting ischemia on stress perfusion scans.32 A more complete understanding of coronary stenosis physiology may require the simultaneous evaluation of coronary flow, coronary pressure, and myocardial perfusion.
Intermediate coronary stenoses are a common angiographic finding; they account for most of the 400 000 angioplasties performed annually.1 9 However, more than 70% of angioplasties performed in the United States are not preceded by a test documenting ischemia,8 and coronary angiography alone does not reliably predict whether an intermediate stenosis causes exercise-induced ischemia. Intracoronary Doppler ultrasound is a widely available, safe, and cost-effective diagnostic modality33 that accurately predicts the presence of ischemia on myocardial perfusion scans. The measurement of coronary flow reserve at the time of angiography documents the physiological significance of intermediate coronary stenoses and therefore provides critical information for determining the need for revascularization.
Study chairperson: Louis I. Heller. Data and Safety Review Committee: D. Douglas Miller, H. Vernon Anderson, and Kinley Larntz. Clinical Coordinating Center and Statistical Group (University of Massachusetts): James Hebert, Thomas Hurley, and Yunsheng Ma. Doppler Core Laboratory (University of Massachusetts Medical Center): Louis I. Heller, Sharon Balcom, and Md. Ashequl Islam. Nuclear Core Laboratory (University of Massachusetts): Jeffrey A. Leppo, Seth T. Dahlberg, Bernard J. Villegas, Mike King, and Brenda McSherry. Angiographic Core Laboratory (Washington Hospital Center): Jeffrey Popma and Allen Merritt. Clinical monitors: Steve Jwanouskos, Collin Sawyer, and Arian Shakeri.
Allegheny General Hospital (Pittsburgh, Pa): James D. Joye. Baylor College of Medicine (Methodist Hospital and Ben Taub Hospital, Houston, Tex): W. Carter Grinstead, Neal S. Kleiman, and Albert E. Raizner. Galicia Medical Group (Wichita, Kan): Robert Kipperman, John Nicholas, and Pat Patterson. St Joseph’s Hospital (Atlanta, Ga): Christopher Cates and Donald Jansen. St. Luke’s Hospital (Milwaukee, Wis): Anita Arnold and Timothy Sommers. University of Massachusetts Medical Center (University Hospital and Berkshire Medical Center), Worcester: Louis I. Heller, Bonnie H. Weiner, Daniel Kusick, Richard Wholey, and Sharon Balcom. West Haven (Conn) Veterans Administration Hospital: Lawrence Deckelbaum and Katherine Rohlfs. Winthrop-University Hospital (Mineola, NY): Richard M. Steingart, Kevin P. Marzo, Anthony T. Gambino, and Mary Ellen Coglianese. Yale University (New Haven, Conn): Michael Cleman, John Setaro, and Diwakar Jain.
This study was supported by a grant from Cardiometrics, Inc. The authors would like to thank Steve Jwanouskos, Mike Billig, and Andy Ford for their expertise, support, and assistance.
↵1 A complete list of Functional Angiometric Correlation with Thallium Scans investigators are listed in the “Appendix.”
- Received November 7, 1996.
- Revision received February 24, 1997.
- Accepted February 28, 1997.
- Copyright © 1997 by American Heart Association
Graves EJ. 1993 Summary: National Hospital Discharge Survey. Atlanta, Ga: Center for Disease Control/National Center for Health Statistics; 1995. US Dept of Health, Education, and Welfare publication.
Topol EJ, Nissen SE. Our preoccupation with coronary luminology: the dissociation between clinical and angiographic findings in ischemic heart disease. Circulation. 1995;92:2333-2342.
Zir LM, Miller SW, Dinsmore RE, Gilbert JP, Harthorne JW. Interobserver variability in coronary angiography. Circulation. 1976;53:627-632.
DeRouen TA, Murray JA, Owen W. Variability in the analysis of coronary angiograms. Circulation. 1977;55:324-328.
Detre K, Wright E, Murphy ML, Takaro T. Observer agreement in the evaluation of coronary angiograms. Circulation. 1975;52:979-986.
Topol EJ, Ellis SG, Cosgrove DM, Bates ER, Muller DW, Schork NJ, Schork MA, Loop FD. Analysis of coronary angioplasty practice in the United States with an insurance-claims database. Circulation. 1993;87:1489-1497.
Rensing BJ, Hermans MD, Deckers JW, De Feyter PJ, Tijssen JGP, Serruys PW. Lumen narrowing after percutaneous transluminal coronary balloon angioplasty nears a Gaussian distribution: a quantitative angiographic study in 1,445 successfully dilated lesions. J Am Coll Cardiol. 1992;19:939-945.
Doucette JW, Corl PD, Payne HM, Flynn AE, Goto M, Nassi M, Segal J. Validation of a Doppler guide wire for intravascular measurement of coronary artery flow velocity. Circulation. 1992;85:1899-1911.
Estes EH. Electrocardiography and vectorcardiography. In: Hurst JW, Logue RB, eds. The Heart. 3rd ed. New York, NY: McGraw-Hill; 1974: chap 21.
Czernin J, Muller P, Chan S, Brunken RC, Porenta G, Krivokapich J, Chen K, Chan A, Phelps ME, Schelbert HR. Influence of age and hemodynamics on myocardial blood flow and flow reserve. Circulation. 1993;88:62-69.
Nichols WW, O’Rourke MF. The nature of flow of a fluid. In: McDonald’s Blood Flow in Arteries. 3rd ed. Philadelphia, Pa: Lea & Febiger; 1990: chap 2.
Mancini G, Simon S, McGillem M, LeFree M, Friedman H, Vogel R. Automated quantitative coronary angiography: morphologic and physiologic validation of a rapid digital angiographic method. Circulation. 1987;75:452-460.
Fleiss JL. Statistical Methods for Rates and Proportions. 2nd ed. New York, NY: John Wiley & Sons; 1981.
Snedecor GW, Cochran WG. Statistical Methods. 8th ed. Ames, Iowa: Iowa State University Press; 1989.
Heller LI, Lewis BS, Silver KH, Donohue TJ, Weiner BH, Kern MJ, Kraft PL, Balcom SJ, Deckelbaum LI, for the National FloWire Registry. Proximal coronary flow reserve does not reflect coronary flow reserve distal to significant stenoses. Circulation. 1994;90(part 2):I-227. Abstract.
Legrand V, Mancini GB, Bates ER, Hodgson JM, Gross MD, Vogel RA. Comparative study of coronary flow reserve, coronary anatomy and results of radionuclide exercise tests in patients with coronary artery disease. J Am Coll Cardiol. 1984;8:1022-1032.
Zijlstra F, Fioretti P, Reiber JH, Serruys PW. Which cineangiographically assessed anatomic variable correlates best with functional measurements of stenosis severity? A comparison of quantitative analysis of the coronary cineangiogram with measured coronary flow reserve and exercise/redistribution thallium-201 scintigraphy. J Am Coll Cardiol. 1988;12:686-691.
Miller DD, Donohue TJ, Younis LT, Bach RG, Aguirre FV, Wittry MD, Goodgold HM, Chaitment BR, Kern MJ. Correlation of pharmacological 99mTc-sestamibi myocardial perfusion imaging with poststenotic coronary flow reserve in patients with angiographically intermediate coronary artery stenoses. Circulation. 1994;89:2150-2160.
Joye JD, Cates CU, Farah T, Knopf W, Soukas PA, Petruolo SM, Donohue BC, Lasorda D. Cost analysis of intracoronary Doppler determination of lesion significance: preliminary results of the PEACH study. J Inv Cardiol. 1994;7:22A. Abstract.