Factors Associated with Noninvasive Coronary Flow Reserve in Severe Aortic Stenosis


Coronary flow reserve (CFR) is progressively impaired with aortic stenosis (AS) severity. However, there is a broad range of CFR in patients with severe AS, and the factors responsible for this variability are weakly characterized. The aim of this study was to assess the correlates of noninvasive CFR in patients with severe AS (≤1 cm 2 or ≤0.6 cm 2 /m 2 ) and preserved left ventricular (LV) ejection fractions (LVEFs) (>50%).


Sixty-six consecutive patients (mean age, 74 ± 11 years; 31 women; mean LVEF, 69 ± 10%) with isolated severe AS (mean, 0.75 ± 0.2 cm 2 and 0.42 ± 0.1 cm 2 /m 2 ), without coronary artery disease, underwent prospectively Doppler transthoracic echocardiography including CFR measurement in the distal part of the left anterior descending coronary artery (LAD) with intravenous adenosine infusion (140 μg/kg/min over 2 min). CFR was defined as hyperemic peak LAD flow velocity divided by baseline flow velocity. Twenty controls matched for age and gender served as a comparative group. Plasma N-terminal pro–brain natriuretic peptide (NT-proBNP) was also assessed.


Compared with controls, patients with AS had higher baseline LAD flow velocities (36 ± 11 vs 27 ± 6 cm/sec, P < 0.01), lower hyperemic LAD flow velocities (80 ± 20 vs 89 ± 18 cm/sec, P = .09), and consequently lower CFR (2.3 ± 0.7 vs 3.3 ± 0.7, P < .01). In patients with AS, there were significant inverse correlations between CFR and age, E/e′, indexed LV mass, NT-proBNP, pulmonary artery systolic pressure (PASP), baseline LV rate-pressure product, heart rate, and indexed left atrial volume and a significant positive correlation between CFR and LVEF (all P values < .05). Furthermore, compared with patients with asymptomatic AS ( n = 22), those with symptomatic AS had more severely impaired CFR (2.15 ± 0.6 vs 2.7 ± 0.65), and higher NT-proBNP values (all P values < .05). In multivariate analysis, NT-proBNP, PASP, and LV rate-pressure product were the main independent correlates of CFR (all P values ≤ .01), and PASP was independently predicted by E/e′ and indexed left atrial volume (all P values < .01).


In patients with severe AS and preserved LVEFs, there is a relatively broad range of CFR values. CFR is more severely impaired in patients with symptomatic AS and is mainly linked with NT-proBNP, a surrogate of increased LV wall stress, workload as measured by LV rate-pressure product, and PASP.

Aortic stenosis (AS) is characterized by a progressive increase of left ventricular (LV) afterload. LV hypertrophy (LVH) is an adaptive response that attempts to reduce wall stress. However, the development of LVH in AS is associated with abnormalities in coronary microcirculatory function. In addition to abnormalities of the resting coronary flow, coronary flow reserve (CFR) is reduced in AS despite angiographically normal coronary arteries, and the reduction of CFR is a determinant of LV ischemia, which portends a poorer prognosis in this setting. Indeed, there is a progressive impairment of CFR with AS severity with a broad range of CFR, even in patients with severe AS. However, the factors responsible for this variability are weakly characterized. Although brain natriuretic peptide (BNP) and N-terminal pro-BNP (NT-proBNP) provide prognostic information beyond clinical and echographic evaluation in AS, the relationship between this important biochemical factor and CFR has never been assessed. Furthermore, LV diastolic dysfunction, which occurs earlier in the course of AS, has important prognostic implications in patients with AS and preserved LV ejection fractions (LVEFs), and its relationship with CFR remains unclear. Doppler transthoracic echocardiography allows the noninvasive evaluation of CFR in various settings with a high success rate, particularly in the left anterior descending coronary artery (LAD). Furthermore, in a stable hemodynamic situation and normal angiographic coronary arteries, CFR is a surrogate for the coronary microcirculation examination. Therefore, our objective was to assess the correlates of noninvasive CFR in patients with severe AS and preserved LVEFs, without coronary artery disease. In view of the observations noted above, we hypothesized that in addition to LVH, impairment of CFR in such patients would be associated with (1) NT-proBNP, a marker of the hemodynamic burden resulting from AS; (2) diastolic dysfunction; and (3) pulmonary artery systolic pressure (PASP).


Between March 1, 2009, and October 30, 2011, 66 consecutive patients with severe AS (≤1 cm 2 or ≤0.6 cm 2 /m 2 ) and preserved LVEFs (≥50%) were prospectively enrolled in this single-center study irrespective of symptom status. They underwent comprehensive Doppler transthoracic echocardiography and noninvasive assessment of CFR in the distal part of the LAD during the same exam. Patients with significant concomitant valvular disease, histories of myocardial infarction, wall motion abnormalities, severe pulmonary disease, and contraindications to adenosine were excluded ( n = 8). All patients underwent coronary angiography. Patients with concomitant coronary diameter stenosis ≥50% were also excluded ( n = 11). Clinical and demographic variables and plasma NT-proBNP were also prospectively assessed. Natriuretic peptide was measured using an immunologic module, an automated Cobas e601 (Roche Diagnostics GmbH, Mannheim, Germany). Venous blood samples were drawn from antecubital veins in patients at rest, the same day as CFR in most cases and within 1 day in the other cases. The presence or absence of symptoms was assessed by two independent cardiologists blinded to patients’ echocardiographic and biochemical data. Patients classified as asymptomatic had to be free of shortness of breath, angina, dizziness, and syncope with exertion. During the same period, 20 normal subjects matched for age and gender, without histories of cardiovascular disease and with normal results on rest electrocardiography and echocardiography, served as a comparative group. All patients gave inform consent to the protocol.

Doppler Transthoracic Echocardiography

Comprehensive echocardiography and CFR measurement were performed using the same commercially available machine (Vivid E9; GE Healthcare, Milwaukee, WI) with an M5S probe. Heart rate and blood pressure were recorded at the same time as echocardiography. All patients were in a stable hemodynamic situation. Aortic valve area was measured using the continuity equation according to current guidelines and indexed to body surface area, which was measured as previously described. The transaortic mean pressure gradient was obtained according to the Bernoulli equation. The peak aortic jet velocity was measured using continuous-wave Doppler from multiple acoustic windows to obtain the highest velocity. The time-velocity integral of the LV outflow tract was measured in the apical five-chamber view, and stroke volume was calculated using LV outflow tract area measured from the parasternal long-axis view. LV end-diastolic and end-systolic volumes and left atrial volume (LAV) were measured from the four-chamber and apical two-chamber views. LVEF was measured according to the modified Simpson’s rule, and LV mass index was measured according to the American Society of Echocardiography formula. Conventional Doppler parameters were also measured according to a standardized examination, and the final value was an average of three cardiac beats: early (E) and late (A) diastolic transmitral flow velocity and E deceleration time; average of the septal and lateral annular mitral early diastolic (e′), late diastolic (a′), and systolic (S) spectral tissue Doppler velocities; and the E/e′ ratio. PASP was calculated as previously described, from the tricuspid regurgitant peak jet velocity, adding right atrial pressure (from respiratory variation of inferior vena cava diameter). LV rate-pressure product (LVRPP), an index of external cardiac work, was calculated as (systolic blood pressure + transaortic peak pressure gradient) × heart rate (mm Hg · beats/min). All echocardiograms including CFR measurements were digitized online on hard disks for subsequent offline analysis, with the support of a dedicated software package (EchoPAC 7 version 108 for PC; GE Healthcare) by one other experienced observer blinded to patient data.


Noninvasive CFR was assessed as previously described, using intravenous adenosine infusion (140 μg/kg/min over 2 min). Briefly, the mid distal part of the LAD was studied using the M5S probe, and the artery was visualized using color Doppler flow mapping guidance, in the modified parasternal view. For color Doppler echocardiography, the velocity range was defined from 12 to 19 cm/sec. Blood flow velocity was measured by pulsed-wave Doppler echocardiography, using a sample volume of 3 to 4 mm, placed on the color signal in the distal LAD. The ultrasound beam direction was aligned as closely as possible with the distal LAD flow. No angle correction was performed for the study, given that CFR is the ratio between hyperemic and baseline flow velocities, and it is not affected by the actual flow velocity. However, the angle was kept as small as possible. CFR was calculated as the ratio of hyperemic to basal peak diastolic flow velocity. Blood flow velocity measurements were performed offline by an experienced investigator blinded to patient data, by contouring the spectral Doppler signals, using the software package of the ultrasound system cited above. Furthermore, diastolic perfusion time (DPT) was measured at baseline and during hyperemia and was defined as (duration of the diastolic component of the coronary flow × heart rate) (sec/min). Final values of flow velocity represented an average of three cardiac cycles. An intravenous contrast agent (SonoVue; Bracco, Milan, Italy) (one bolus of 0.1 mL repeated if necessary) was used in 16% of patients to improve visualization of the color Doppler signal and/or to obtain clear spectral Doppler signal in the LAD. Heart rate was monitored continuously during the patient examination, and blood pressure was recorded at baseline and during hyperemia using an automatic arm sphygmomanometer. Intraobserver and interobserver variability of CFR was evaluated from 10 randomly selected patients with AS and calculated as the absolute difference divided by the average of the two observations times 100. Mean intraobserver variability expressed was 3.6 ± 3.3%. Mean interobserver variability was 4.1 ± 4%.

Statistical Analysis

Continuous variables are expressed as mean ± SD and categorical variables as percentages. When a variable was not normally distributed, a logarithmic transformation was performed (NT-proBNP). Unpaired or paired Student’s t tests and χ 2 tests (or Fisher’s exact tests as appropriate) were performed to assess differences according to the variables tested. The relationship between CFR and several parameters was tested with linear and nonlinear correlations, and the best fit was retained. A receiver operating characteristic curve was created to assess the best cutoff of CFR and hyperemic LAD flow velocity associated with symptomatic status. A stepwise multiple logistic regression analysis was performed to identify independent predictors of CFR and baseline and hyperemic LAD flow velocities. Variables related to the dependent variable in univariate analysis ( P < .10) were included in the multivariate model. However, because of the relatively large number of variables significantly associated with CFR and baseline LAD flow velocity in univariate analysis compared with the number of observations, the choice of the variables entered in the multivariate model was based on the following principles: the strength of association with the dependent variable in univariate analysis, an attempt to avoid collinearity with another variable, and eventually its background value on the basis of previous studies in AS. Statistical analysis was performed using MedCalc for Windows version (MedCalc Software, Mariakerke, Belgium). P values < .05 were considered significant.


Baseline characteristics ( Table 1 ) were similar between AS and controls, as well as LVEF and LV volumes (all P values = NS). Markers of diastolic function were significantly impaired, and indexed LV mass, LVRPP, and PASP were significantly higher in patients with AS than in controls (all P values < .01; Table 2 ). The etiology of AS was degenerative in most cases ( n = 52) and bicuspid aortic valve in the other cases ( n = 14). Forty-four patients (67%) were symptomatic (exertional dyspnea, New York Heart Association class II in 20, New York Heart Association class III in 15, syncope in three, dizziness in two, and angina pectoris in 10), and 22 (33%) were asymptomatic. Adenosine tests were well tolerated, and no serious adverse events occurred during adenosine infusion ( Table 3 ). An example of CFR in a patient with AS is depicted in Figure 1 . Compared with controls, patients with AS had higher baseline LAD flow velocities (36 ± 11 vs 27 ± 6 cm/sec, P < .01), lower hyperemic LAD flow velocities (80 ± 21 vs 89 ± 18 cm/sec, P = .09), and consequently lower CFR (2.3 ± 0.7 vs 3.3 ± 0.7, P < .01). In patients with AS, there was a relatively broad range of CFR (from 1.3 to 4.4; Figure 2 ). Compared with asymptomatic patients ( n = 22), symptomatic patients ( n = 44) had higher end-systolic volumes, PASP, and indexed LAVs and lower LVEFs and E deceleration times (all P < .05), higher NT-proBNP values ( P < .01), and more impaired CFR (2.15 ± 0.6 vs 2.7 ± 0.65, P < .01; Figure 3 ), because of reduced vasodilating capacity (peak hyperemic LAD flow velocity, 75 ± 18 vs 92 ± 19 cm/sec, P < .01; baseline LAD flow velocity, 36 ± 11 vs 36 ± 12 cm/sec, P = NS). On receiver operating characteristic curve analysis, the best cutoff of CFR associated with symptomatic status was 2.15, with sensitivity of 66% and specificity of 86% and an area under the curve of 0.75 ( P < .01; Figure 4 A). In comparison, the best cutoff hyperemic LAD flow velocity associated with symptoms was 86 cm/sec, with sensitivity of 77% and specificity of 68% and an area under the curve of 0.74 ( P < .01; Figure 4 B). In patients with AS, significant correlations were observed between CFR and clinical, hemodynamic, and echographic parameters and NT-proBNP. Table 4 and Figure 5 summarize these correlations, relatively weak for some factors, between CFR and age, heart rate, resting DPT, LVRPP, indexed LV mass, LVEF, E/e′ ratio, LAV, PASP, and NT-proBNP. Such significant correlations were not found in the control group (age, r = 0.18; heart rate, r = −0.30; resting DPT, r = −0.30; LVRPP, r = 0.10; indexed LV mass, r = −0.30; LVEF, r = 0.10; E/e′, r = −0.29; LAV, r = −0.20; and PASP, r = −0.06; all P values = NS). Furthermore, CFR did not differ significantly in patients with AS with and without vascular risk factors and according to gender ( P = NS for all). According to AS etiology, CFR was significantly higher in patients with bicuspid compared with degenerative AS (2.8 ± 0.6 vs 2.2 ± 0.6, P < .01). However, after adjustment for age (bicuspid aortic valve, 61 ± 11 years; degenerative AS, 77 ± 9 years; P < .01) and diastolic dysfunction parameters, CFR was not independently associated with bicuspid etiology. Resting and hyperemic LAD flow velocities were also correlated with several parameters in patients with AS ( Table 4 ).

Table 1

Baseline characteristics

Variable Patients with AS
( n = 66)
( n = 20)
Age (y) 74 ± 11 70 ± 10
Women 34 (52%) 10 (50%)
Body mass index (kg/m 2 ) 26.5 ± 5.4 24.8 ± 4.5
Body surface area (m 2 ) 1.8 ± 0.2 1.8 ± 0.2
Hypertension 43 (65%) 8 (40%)
Diabetes 13 (20%) 2 (10%)
Smoking 9 (14%) 2 (10%)
Dyslipidemia 39 (59%) 8 (40%)
Log NT-proBNP 2.8 ± 0.68
Heart rate (beats/min) 67 ± 11 65 ± 10
Mean blood pressure (mm Hg) 90 ± 13 92 ± 11

Data are expressed as mean ± SD or as number (percentage). P = NS for all variables.

Table 2

Doppler echocardiographic variables

Variable Patients with AS Controls
LVEF (%) 69 ± 10 70 ± 5
End-diastolic volume (mL) 97 ± 28 97 ± 20
End-systolic volume (mL) 30 ± 15 29 ± 6
Aortic valve area (cm 2 ) 0.75 ± 0.2
Indexed aortic valve area (cm 2 /m 2 ) 0.42 ± 0.1
Mean gradient (mm Hg) 46 ± 15
Peak aortic jet velocity (m/sec) 4.3 ± 0.6
Stroke volume (mL/m 2 ) 41 ± 8 39 ± 8
LVRPP (mm Hg · beats/min) 14,710 ± 3,741 8,364 ± 1,448
LV mass (g/m 2 ) 111 ± 28 75 ± 10
LAV (mL/m 2 ) 33 ± 11 22 ± 5
E (cm/sec) 83 ± 30 67 ± 17
A (cm/sec) 96 ± 32 71 ± 20
E/A ratio 0.9 ± 0.4 1 ± 0.4
E deceleration time (msec) 259 ± 86 213 ± 47
e′ (cm/sec) 5.9 ± 2.4 7.5 ± 1.5
a′ (cm/sec) 8.8 ± 2.4 9.2 ± 1.9
S (cm/sec) 6.2 ± 1.5 8 ± 1.5
E/e′ ratio 15.7 ± 7.3 9.5 ± 3.2
PASP (mm Hg) 39 ± 9 30 ± 5

Data are expressed as mean ± SD.

P ≤ .01.

Table 3

Coronary flow velocity and hemodynamic data

Variable Patients with AS Controls
Baseline Hyperemia Baseline Hyperemia
Heart rate (beats/min) 67 ± 11 72 ± 12 65 ± 10 77 ± 14
Systolic blood pressure (mm Hg) 137 ± 20 134 ± 21 137 ± 18 136 ± 11
Diastolic blood pressure (mm Hg) 67 ± 11 65 ± 15 68 ± 11 67 ± 11
LAD flow velocity (cm/sec) 36 ± 11 80 ± 20 27 ± 6 89 ± 18

Data are expressed as mean ± SD.

P < .01 versus patients with AS.

P < .01 versus baseline.

Figure 1

Example of an asymptomatic patient with preserved CFR despite severe AS. ( Top left ) Example of continuous-wave Doppler in the right parasternal view, showing a peak aortic jet velocity at 5.2 m/sec and a mean gradient at 63 mm Hg (calculated aortic valve area, 0.7 cm 2 ). ( Top right ) a sample of a color Doppler LAD flow is depicted. ( Bottom ) CFR: ( left ) coronary flow velocity at baseline and ( right ) coronary flow velocity during hyperemia. CFR = 3.2.

Figure 2

Comparison of CFR between patients with severe AS and controls. Mean CFR ( box ), standard deviation ( bars ), and plots are depicted in each group. The control group is shown in red and the AS group in yellow.

Jun 7, 2018 | Posted by in CARDIOLOGY | Comments Off on Factors Associated with Noninvasive Coronary Flow Reserve in Severe Aortic Stenosis

Full access? Get Clinical Tree

Get Clinical Tree app for offline access