The aim of this study was to test whether acceleration time of systolic coronary flow velocity could contribute to the diagnosis of coronary stenosis in patients with microvascular dysfunction, on the basis of the hypothesis that systolic coronary flow is less influenced by microvascular function because of compressed myocardium.
Coronary flow velocity was assessed in the left anterior descending coronary artery during hyperemia with intravenous adenosine by echocardiography in 502 patients who were scheduled for coronary angiography because of coronary artery disease and significant valvular disease. Coronary flow velocity reserve (CFVR) and the percentage acceleration time (%AT), as the percentage of the time from the beginning to the peak of systolic coronary flow over systolic time during hyperemia, were calculated. The diagnostic ability of CFVR and %AT for angiographic coronary artery stenosis was then analyzed. As invasive substudies, fractional flow reserve and %AT by a dual-sensor (pressure and Doppler velocity) guidewire were measured simultaneously with %AT on transthoracic echocardiography ( n = 14).
Patients with coronary stenosis had significantly lower CFVR (1.7 ± 0.4) and greater %AT (65 ± 9%) compared with those without stenosis (2.6 ± 0.6 and 50 ± 13%, respectively). Percentage acceleration time by Doppler echocardiography was in good agreement with %AT ( r = 0.98) and fractional flow reserve ( r = 0.74) invasively measured by dual-sensor guidewire. Cutoff values of CFVR and %AT were determined as 2.0 and 60% in receiver operating characteristic curve analysis. The sensitivity, specificity, and accuracy of CFVR to detect coronary stenosis were 71.1%, 77.3%, and 75.4%, while those of %AT were 83.4%, 71.8%, and 75.4%, respectively. In addition, %AT provided high accuracy to detect coronary stenosis, especially in patients with previous myocardial infarctions, valvular disease, and left ventricular hypertrophy (81.1%, 84.1%, and 73.4%, respectively).
The %AT of systolic coronary flow velocity is a promising marker to diagnose coronary stenosis in patients with microvascular dysfunction.
Coronary flow velocity reserve (CFVR) by transthoracic echocardiography has been considered a useful diagnostic index for functional and physiologic assessment of coronary circulation. Clinical application of CFVR is limited to patients without microvascular dysfunction because it is altered either by the presence of epicardial coronary artery stenosis or an abnormality of the coronary microcirculation. In the clinical setting, patients with epicardial coronary stenosis commonly have one or more additional factors affecting the microvasculature. Therefore, a noninvasive and physiologic method allowing the assessment of coronary stenosis in patients with microvascular dysfunction is needed.
Previous studies have demonstrated that the acceleration time of systolic flow velocity is proportional to the severity of stenosis in peripheral arteries. However, the relationship between systolic acceleration time and the severity of coronary stenosis has not been investigated. Unlike peripheral arteries, flow velocity in the coronary arteries could be influenced by coronary microvasculature; moreover, microvascular resistance is increased with an increase in the severity of coronary stenosis in patients without collateral flow. In view of the fact that the myocardium thickens in systole, compressed intramyocardial microvessels thus lead to significantly diminished microvascular blood flow. Therefore, epicardial systolic coronary flow would be less influenced by microvascular circulation.
We sought to investigate whether systolic acceleration time of coronary flow velocity could be useful to diagnose coronary artery stenosis in patients with microvascular dysfunction. In the present study, systolic acceleration time was measured during hyperemia to improve the resolution of the spectral tracing.
We enrolled 502 patients who were scheduled for diagnostic coronary angiography from January 1, 2008, to February 28, 2011. They were admitted to our hospital for the assessment of heart disease, including significant valvular disease and coronary artery disease. Comprehensive echocardiographic examinations and coronary flow assessments were performed in all patients <48 hours before coronary angiography. Coronary flow studies by transthoracic echocardiography only were also performed in 10 healthy volunteers with no obvious ischemia as determined by exercise echocardiography (mean age, 35 ± 9 years; seven men). Exclusion criteria were acute coronary syndromes, congestive heart failure or nonsinus rhythm, and contraindication to adenosine (asthma, high-degree atrioventricular block). Prior myocardial infarction (MI) was defined as a history of MI (>1 month). The study protocol was approved by the local ethics committee, and written informed consent was obtained. All patients continued their medications on the day of echocardiographic examination and coronary angiography.
Left ventricular (LV) diameters were measured on two-dimensional images. LV ejection fraction (LVEF) and left atrial volume were calculated using the biplane modified Simpson’s rule using apical four-chamber and two-chamber views. LV mass was calculated using the Devereux formula indexed to body surface area (LV mass index). Diastolic parameters using transmitral flow patterns and mitral annular velocities were also determined. Patients were considered to have ventricular hypertrophy when LV mass index was >150 g/m 2 in men and >120 g/m 2 in women.
Measurement of Coronary Flow Velocity
Coronary flow velocity was recorded using a Vivid 7 system (GE Healthcare, Milwaukee, WI) along with a 4-MHz transducer after an overnight fast and abstention from any beverages containing significant amount of flavonoids for 48 hours to avoid the effects of flavonoids on improving coronary endothelial function. First, we assessed blood flow in the left anterior descending coronary artery (LAD), which appears as a red color signal in the anterior interventricular sulcus during diastole, using Doppler color flow mapping. Second, we positioned the sample volume on the color signal in the distal LAD and measured coronary flow velocities by pulsed-wave Doppler echocardiography.
Doppler flow velocity recording in the LAD was performed at baseline and during hyperemia, which was induced by intravenous adenosine triphosphate administration (0.14 mg/kg/min). Heart rate and blood pressure were monitored continuously during the examination. Coronary flow velocities were measured offline by tracing the spectral Doppler signals (EchoPAC version 6.1; GE Vingmed Ultrasound AS, Horten, Norway). Mean diastolic flow velocities at baseline and hyperemia were averaged over three consecutive cycles. CFVR was calculated as the ratio of hyperemic to basal mean diastolic flow velocities. Systolic coronary flow velocities during hyperemia were assessed offline as well. Acceleration time was determined as the time from the beginning to the peak systolic flow and percentage acceleration time (%AT) was defined as (acceleration time/systolic time) × 100 ( Figure 1 A ).
All patients received an intravenous bolus injection of heparin 3,000 IU and intracoronary isosorbide dinitrate 2 mg before angiography. Quantitative coronary angiographic analysis was performed offline in multiple projections using a guiding catheter to calibrate magnification, as previously described. The analyses were performed by 2 independent observers who were completely blinded to any patient information. Significant coronary stenosis was defined as ≥50% luminal diameter narrowing on the angiogram.
As invasive substudies, %AT was measured using a Doppler flow guidewire (FloWire; Volcano, San Diego, CA) positioned in the distal LAD and compared with simultaneously determined %AT by transthoracic echocardiography in 14 patients ( Figure 1 B). In addition, to assess the effect of adenosine on %AT, we made a comparison of %AT values at rest and during hyperemia using the dual-sensor guidewire recordings.
Fractional flow reserve (FFR) and %AT using a 0.014-inch dual-sensor (pressure and Doppler velocity) guidewire (ComboWire; Volcano Therapeutics, Rancho Cordova, CA) were measured simultaneously with %AT by transthoracic echocardiography. FFR was calculated as dividing the mean distal coronary pressure measured with the dual-sensor guidewire by the mean aortic pressure measured through the guide catheter.
Interobserver and intraobserver variability in coronary flow velocity measurements were determined in 40 randomly selected patients. Interobserver variability was calculated as the standard deviation of the absolute differences between the measurements made by two independent observers who were blinded to patient information. Intraobserver variability was also calculated as the standard deviation of the absolute differences between the first and second measurements (2-week interval) for a single observer. Variability is expressed as a percentage of the average value. The mean absolute differences in %AT and CFVR were 5.1 ± 4.7% and 4.2 ± 3.9% (interobserver) and 4.9 ± 4.7% and 4.0 ± 3.8% (intraobserver), respectively.
All analyses were conducted using SPSS for Windows version 13.0 (SPSS, Inc, Chicago, IL). Continuous data are expressed as mean ± SD and were compared across groups using analysis of variance with Scheffé’s post hoc comparison. Categorical data are presented as absolute values and percentages, and they were compared using χ 2 tests. Receiver operating characteristic (ROC) curves were constructed to evaluate the predictive performance of CFVR and %AT to coronary stenosis in sequential patients enrolled from January 1 to December 31, 2008 (a derivation cohort). The diagnostic accuracy of CFVR and %AT for coronary artery stenosis was analyzed using the determined cutoff value by ROC curve analysis in sequential patients from January 1, 2009, to February 28, 2011 (a test cohort). P values < .05 were considered statistically significant.
A total of 466 of 502 enrolled patients (93%) underwent the study ( Figure 2 ). Twenty-five patients were excluded because of low resolution of systolic flow profile recordings. An additional six were excluded because of retrograde LAD flow, along with five with no coronary signal in anterior interventricular sulcus, of which total LAD occlusion by coronary angiography was found in the former six and two in the latter case. In 466 patients, 212 with previous MIs, 75 with valvular disease (41 with aortic stenosis [AS], seven with aortic regurgitation, 24 with mitral regurgitation, and three with mitral stenosis) and 66 with LV hypertrophy (LVH) were included. With regard to previous MI, culprit vessels were the LAD in 103 patients and the right coronary artery or circumflex branches in 109.
Baseline Characteristics and Echocardiographic Data
Table 1 presents general baseline characteristics and echocardiographic data of the patients with percentage diameter stenosis (%DS) ≥50% and <50%. Patients with %DS ≥ 50% were more likely to have hypertension than those with %DS < 50%. Regarding New York Heart Association classification, there was no patients in class III or IV and no difference in the frequency of class II between two groups. All echocardiographic measurements were similar in the two groups.
|Variable||%DS ≥ 50% |
( n = 147)
|%DS < 50% |
( n = 319)
|Age (y)||69 ± 9||67 ± 12||.800|
|Men||110 (75%)||220 (70%)||.196|
|Diabetes mellitus||57 (39%)||85 (27%)||.008|
|Hypertension||99 (67%)||188 (59%)||.177|
|Dyslipidemia||75 (51%)||138 (44%)||.119|
|Smoking||56 (38%)||82 (26%)||.317|
|NYHA class II||13 (11%)||50 (15%)||.074|
|MI||63 (43%)||149 (48%)||.437|
|Valvular disease||10 (7%)||65 (20%)||.001|
|LVH||16 (11%)||50 (16%)||.208|
|%DS (%)||69 ± 10||26 ± 13||<.001|
|Reference diameter (mm)||3.1 ± 0.6||3.0 ± 0.6||.189|
|Minimum lumen diameter (mm)||1.0 ± 0.4||2.2 ± 0.7||<.001|
|Lesion length (mm)||24.5 ± 12.9||16.9 ± 8.1||.002|
|Multiple lesions||38 (46%)||41 (11%)||<.001|
|LVDd (mm)||46 ± 8||47 ± 7||.099|
|LVDs (mm)||32 ± 9||31 ± 8||.943|
|IVSWT (mm)||10.2 ± 3.9||10.5 ± 5.2||.863|
|PWT (mm)||10.1 ± 5.1||10.0 ± 1.8||.283|
|LVEDV (mL)||87 ± 25||87 ± 33||.937|
|LVESV (mL)||39 ± 16||43 ± 44||.411|
|LVEF (%)||56 ± 10||55 ± 10||.302|
|LVMI (g/m 2 )||120 ± 28||120 ± 31||.996|
|LAVI (mL/m 2 )||33 ± 16||39 ± 28||.139|
Coronary Angiographic Findings
Coronary angiography revealed significant stenosis of the LAD in 147 of 466 patients (32%). The lesion lengths of the group with %DS ≥ 50% were longer than those in the group with %DS < 50% ( Table 1 ). Multiple lesions were found more frequently in the group with %DS ≥ 50% than the group with %DS < 50%. In addition, to assess the effect of lesion characteristics on CFVR and %AT, we compared data of multiple lesions with those of focal lesions in 142 of 466 patients with intermediate stenosis (%DS, 40% to 70%). We found impaired CFVR (2.0 ± 0.5) and %AT (68 ± 10%) in 66 patients with multiple lesions in comparison with 76 patients with focal lesions (2.5 ± 0.8 and 59 ± 10%, P = .02 and P = .004, respectively), while %DS was similar in both groups (multiple vs focal, 59 ± 10% vs 57 ± 9%).
To determine the reliability of transthoracic echocardiography in %AT measurements, we additionally performed intracoronary flow studies in 14 patients with various levels of severity of coronary stenosis. The results revealed that %AT measured by transthoracic echocardiography is indeed consistent with invasively measured %AT ( r = 0.98, P < .0001; Figure 3 A ). The systolic flow velocity analysis using dual-sensor guidewire recordings revealed that resting %AT (57 ± 10%) did not change significantly during hyperemia (56 ± 11%), while systolic peak velocity changed from 13 ± 9 to 27 ± 14 cm/sec.
In addition, %AT decreased in inverse proportion to FFR ( r = 0.74, P = .014; Figure 3 B). Consequently, we assessed the feasibility of %AT by transthoracic echocardiography to diagnose coronary stenosis in the clinical settings.
Coronary Flow Velocity Measurements
Analysis of coronary flow velocity measurements in 466 patients revealed severely reduced CFVR (1.7 ± 0.4) and prolonged %AT (65 ± 9%) in patients with %DS ≥ 50%. Patients with %DS < 50% had higher CFVR (2.6 ± 0.6) and shorter %AT (50 ± 13%) compared with those with %DS < 50% ( Figure 4 ). Percentage acceleration time was proportional to %DS as well ( Figure 5 ).
Moreover, although healthy subjects showed higher CFVR (3.3 ± 0.7) than those with %DS < 50%, there was no significant difference in %AT (in healthy subjects, 48 ± 10%). There were no significant differences in blood pressure and heart rate at baseline and during hyperemia ( Table 2 ). To assess the effect of LV systolic function on %AT, we divided the 466 patients into two groups: 246 with LVEFs > 55% and 220 with LVEFs ≤ 55%. These two groups had similar %DS (39.6 ± 23.4% and 38.1 ± 21.9%, respectively, P = .48). However, the patients with LVEFs ≤ 55% had lower CFVR (2.3 ± 0.8) than those with LVEFs > 55% (2.5 ± 0.8, P = .005), whereas no difference in %AT (LVEF > 55% vs ≤ 55%, 54 ± 13% vs 53 ± 14%; P = .51).
|Variable||%DS ≥ 50%||%DS < 50%|
|Heart rate (beats/min)||65 ± 11||66 ± 12||69 ± 14||69 ± 13|
|Systolic BP (mm Hg)||117 ± 19||114 ± 19||116 ± 19||111 ± 20|
|Diastolic BP (mm Hg)||59 ± 11||56 ± 13||61 ± 13||58 ± 12|
|MDV (cm/sec)||25 ± 14||41 ± 22||22 ± 8||57 ± 19|
|CFVR||1.7 ± 0.4 ∗||2.6 ± 0.6|
|%AT||65 ± 9 ∗||50 ± 13|