Download Echocardiographic Evaluation Before Bidirectional Glenn Operation

Echocardiographic Evaluation Before Bidirectional Glenn
Operation in Functional Single-Ventricle Heart Disease
Comparison to Catheter Angiography
Kenan W.D. Stern, MD; Doff B. McElhinney, MD; Kimberlee Gauvreau, ScD;
Tal Geva, MD; David W. Brown, MD
Downloaded from http://circimaging.ahajournals.org/ by guest on May 11, 2017
Background—Cardiac catheterization is routinely performed in patients with single ventricle before bidirectional Glenn
operation (BDG). There is interest in noninvasive evaluation alone before BDG, but concern for echocardiography
successfully imaging the relevant anatomy persists. We evaluated the accuracy of echocardiographic imaging of
vascular anatomy.
Methods and Results—Diagnostic images of 130 patients who had echocardiography and catheterization before BDG were
reviewed; diameters of the pulmonary arteries (PAs) and aortic arch were measured, and stenoses were recorded. Patient and
procedural factors associated with echocardiographic imaging were analyzed. Median age at echocardiography was 4 months;
the most common diagnosis was hypoplastic left heart syndrome (55%). The left PA was imaged by echocardiography in 83
patients (64%), with 4 of 21 stenoses (19%) diagnosed by catheterization identified; similarly, the right PA was imaged in 81
(62%), and 3 of 17 stenoses (18%) were identified. The distal aortic arch was visualized in 104 (80%), with successful
identification of 21 of 27 (78%) of coarctations diagnosed by catheterization. Complete vascular echocardiography
(visualization of PAs and aortic arch) occurred in 43% and was not obtained more frequently with sedation.
Conclusions—In a large cohort of patients presenting for BDG, evaluation by echocardiography frequently failed to image the
PAs and missed the majority of PA stenoses. Sedation did not appear to improve the performance of echocardiography for
evaluation of the PAs. Echocardiography cannot be relied on as the sole investigation before BDG. (Circ Cardiovasc
Imaging. 2011;4:498-505.)
Key Words: congenital heart disease 䡲 echocardiography 䡲 single ventricle 䡲 superior cavopulmonary connection
P
atients with congenital heart disease with single-ventricle
physiology typically undergo staged surgical palliation culminating in the Fontan procedure. Bidirectional Glenn operation
(BDG) is often performed in infancy as an intermediate procedure, and results in passive flow of venous return from the upper
body through the pulmonary vascular bed. Impedance to blood
flow through the pulmonary circulation leads to poor postoperative outcomes after BDG.1–3 In addition, the presence of aortic
arch obstruction is a risk factor for late mortality after stage I
Norwood operation.4 –5 Therefore, adequate evaluation of this
vascular anatomy is required before surgery.
ever, with marked improvement in noninvasive imaging
modalities including echocardiography and cardiac magnetic
resonance, as well as the technical difficulty, expense, and
ionizing radiation exposure of catheterization, there has been
increasing interest in noninvasive evaluation alone.7,8 Retrospective studies have demonstrated that patients are rarely
excluded from BDG on the basis of findings at catheterization, and that routine catheterization after an unremarkable
echo is of limited clinical utility.7,9 A recent randomized,
prospective study of cardiac magnetic resonance versus
catheterization before BDG showed similar outcomes, fewer
minor complications, and lower hospital charges with cardiac
magnetic resonance.10 However, cardiac magnetic resonance
is resource-intensive, technically challenging, and requires
general anesthesia in this patient population. These findings
raise the question of whether echo may be used as the sole
preoperative imaging modality.
We therefore sought to evaluate the accuracy and reliability of echo in this setting as compared with the gold standard
Clinical Perspective on p 505
Cardiac catheterization has been performed routinely in the
evaluation of patients presenting for BDG to assess for
hemodynamic and anatomic issues that might complicate or
exclude patients from BDG, including the assessment of the
pulmonary vascular anatomy, and to intervene on such
conditions with transcatheter therapy if appropriate.6 How-
Received January 3, 2011; accepted June 21, 2011.
From the Department of Cardiology, Children’s Hospital Boston and Department of Pediatrics, Harvard Medical School, Boston, MA.
Guest Editor for this article was Craig E. Fleishman, MD.
Correspondence to David W. Brown, MD, Department of Cardiology, Children’s Hospital, 300 Longwood Ave, Boston, MA 02115. E-mail
[email protected]
© 2011 American Heart Association, Inc.
Circ Cardiovasc Imaging is available at http://circimaging.ahajournals.org
498
DOI: 10.1161/CIRCIMAGING.110.963280
Stern et al
Echo Before Glenn Operation
499
Downloaded from http://circimaging.ahajournals.org/ by guest on May 11, 2017
Figure 1. Location of pulmonary artery measurements on echocardiography. A, Right pulmonary artery behind superior vena cava and
central pulmonary artery behind ascending aorta. B, Left pulmonary artery anterior to descending aorta.
of catheterization, with particular attention to vascular anatomy and dimensions, and to evaluate the relationship between
patient or procedural factors and echo performance.
Methods
Patients who underwent BDG at Children’s Hospital Boston from
January 2004 to April 2008 were identified by search of the
Cardiovascular Program database. Patients were included for analysis if they had both echo and catheterization performed before BDG
and had images available for review in our digital imaging database.
Patient demographic, clinical, and procedural variables were collected for analysis.
Pre-BDG studies were defined as those echos and catheterizations
performed ⱕ60 days of BDG. Echo did not have to precede
catheterization to be considered the pre-BDG study. For patients
with multiple echos during this period, the most complete study was
selected as the pre-BDG study. The study was approved by the
Department of Cardiology Scientific Review Committee and by the
Children’s Hospital Boston Institutional Review Board, with waiver
of informed consent.
Echocardiography
Digital images from pre-BDG echos were reviewed by an investigator blinded to the results of other imaging studies or intraoperative
findings (K.S.). Blinding was achieved by the institution of a
temporal gap of approximately 6 months between review of catheterization angiograms and echo images. Intraoperative findings were
reviewed last after having collected catheterization and echo data.
Images were evaluated for visualization of the structure of interest,
and measurements of vascular structures were made to the nearest
0.1 mm. Branch pulmonary artery (PA) z-scores were recorded. A
second investigator (D.W.B.) blinded to the results of other
imaging studies repeated measurements and assessment of central
PA distortion on a randomly selected 10% sample to assess
interobserver variability. Earlier echos in the sample were obtained with HP SONOS 7500 and 5500 ultrasound scanners
(Hewlett-Packard Co, Palo Alto, CA), and later studies were
obtained with the Philips IE33 (Philips Medical Systems, Andover, MA). Images were reviewed on HeartSuite Vericis image
review software (Emageon, Birmingham, AL).
Complete vascular echo was defined as adequate visualization of
the central PA, mediastinal branch PAs, distal aortic arch and
proximal descending aorta.
Visualization of the central PA and mediastinal branch PAs was
typically from parasternal imaging views (Figure 1). The mediastinal
right PA (RPA) was measured at the level of the right superior vena
cava, the central PA in the midline typically behind the ascending
aorta, and the mediastinal left PA (LPA) at the level of the
descending aorta. The central PA was assessed qualitatively for
degree of distortion, which was defined as abnormal tortuosity, and
characterized as undistorted, minor, or major distortion. If stenosis
was present in the central or branch PAs, the narrowest diameter and
the adjacent unobstructed segment were measured and percent
stenosis was calculated. Stenoses were categorized as no significant
stenosis (⬍25%) or mild (25% to 50%), moderate (51% to 75%), or
severe (⬎75%), in accordance with the previously published
criteria.9
Echo assessment of the aortic arch was typically obtained from
high parasternal or suprasternal notch imaging views. Measurements
were made at the narrowest portion of the aortic isthmus and at the
adjacent unobstructed portion of the proximal descending aorta.
Doppler gradients corrected for proximal velocity across the distal
aortic arch were categorized as unobstructed (⬍10 mm Hg) or
obstructed (ⱖ10 mm Hg).
Catheter Angiography
Pre-BDG catheterization digital angiograms were reviewed by an
investigator blinded to the results of the echo images (K.S.).
Angiograms were assessed for visualization of the structures of
interest, and vascular measurements were made to the nearest
0.1 mm with the use of electronic calipers on HeartSuite Vericis
image review software version 6.20.1 (Emageon, Birmingham, AL).
For comparison of branch PA hypoplasia with echo, branch PA
z-scores were also calculated for angiographic measurements. A
second investigator (D.W.B.), blinded to the results of other studies,
performed repeat measurements and assessment of central PA
distortion of catheter angiograms on a randomly selected 10%
sample to assess for interobserver variability.
Qualitative assessment of the degree of distortion of the central PA
was reported as undistorted, minor, or major distortion (Figure 2).
Measurements of PA angiograms were preferentially made in the AP
projection with digital calibration. Measurements were performed at
the same location as on echo. For stenoses, the narrowest diameter
was measured as well as the immediately adjacent nonstenotic
segment. For the distal aortic arch, the absolute diameter of the
narrowest segment of the distal arch was measured preferentially
through the lateral projection, with measurement of the adjacent
unobstructed proximal descending aorta. For patients who underwent
intervention on the PAs or the aortic arch, catheterization measurements were made either before or after intervention, depending on
whether the pre-BDG echo was a precatheterization or postcatheterization study. For example, a patient undergoing balloon dilation of
the RPA with a postcatheterization echo had postintervention catheterization measurements used for comparison with echo and was not
excluded from analysis.
500
Circ Cardiovasc Imaging
September 2011
Figure 2. Pulmonary artery distortion assessment. Catheter angiograms demonstrating examples of A, no distortion; B, mild distortion;
and C, major distortion of the central pulmonary artery.
Statistical Analysis
Downloaded from http://circimaging.ahajournals.org/ by guest on May 11, 2017
Patient characteristics were compared for subjects with complete
versus incomplete vascular echos using the Wilcoxon rank sum test
for continuous variables and Fisher exact test for categorical variables. Categorical variables measured by both echo and catheterization were cross-tabulated, and percent agreement of echo with
catheterization was calculated. For continuous variables measured by
both techniques, analysis was performed only on patients who had
echo and catheterization performed not more than 14 days apart; a
95% confidence interval for the mean difference was calculated, and
statistical significance was assessed using the paired t test.
Analyses were performed overall and then separately for sedated
and unsedated echos. Differences were illustrated using the
Bland-Altman method. Associations between echo and catheterization measurements were also quantified using the Pearson
correlation coefficient. For a random sample of 10% of study
subjects, interobserver variability was assessed for both echo and
catheterization measurements by calculating a 95% confidence
interval for the mean difference between observers and by the
paired t test. Analyses were performed using Stata version 10
(StataCorp LP, College Station, TX).
Results
A total of 207 patients underwent BDG at Children’s Hospital
Boston from January 2004 through April 2008. From this
population, 130 patients who had both pre-BDG echo and
catheterization performed within 60 days of surgery were
included. The majority of exclusions were for lack of digital
catheterization (n⫽38) or echo (n⫽28) images.
Baseline patient characteristics are shown in Table 1.
Median age and weight at time of echo were 4 months and 5.7
kg, respectively; median time between pre-BDG echo and
catheterization was 1 day. The time difference from echo to
catheterization ranged from 55 days before to 57 days after;
however, 116 studies took place within 14 days of each other,
with the majority (n⫽100) on either the same or consecutive
days. Most of the echos (n⫽102) were performed before the
catheterization.
In a sample of 10% of study patients, there were no
significant differences between reviewers for either echo
or catheterization measurements of the PAs, distal aortic
arch, or proximal descending aorta. Agreement on central
PA distortion was 100%.
Echo Completeness
A complete vascular echo was obtained in 56 patients (43%).
No patient characteristics were associated with the likelihood
of obtaining a complete vascular echo, including anatomic
diagnosis and prior interventions. Timing of echo relative to
catheterization was not significantly related to obtainment of
a complete vascular echo (Table 2).
Echo Assessment of Branch PAs
The LPA was imaged by echo in 83 (64%) patients. LPA
stenosis was present by catheterization in 21 patients (16%):
mild in 12 (9%), moderate in 8 (6%), no severe stenoses, and
1 case of LPA atresia (1%). For stenoses, 5 of 21 cases were
identified by echo, 4 of which were accurately categorized by
severity. For mild LPA stenosis, 2 of 12 cases were correctly
Table 1.
Patient Characteristics (nⴝ130)
Variable
No. (%) or
Median (Range)
Sex
Male
71 (55%)
Female
59 (45%)
Age, mo*
4 (1 to 61)
Height, cm, n⫽84
62 (47 to 82)
Weight, kg
5.7 (2.4 to 19.7)
Time from echo to catheterization, d
1 (⫺57 to 55)
Anatomic type of single ventricle
Hypoplastic left heart syndrome
71 (55%)
Complex 2-ventricle
14 (11%)
Double-inlet left ventricle
11 (8%)
Single right ventricle
11 (8%)
Tricuspid atresia
9 (7%)
Unbalanced AV canal
8 (6%)
Pulmonary atresia with intact ventricular septum
6 (5%)
Previous surgery
Stage I/RVPA conduit
46 (35%)
Stage I/modified BTS
37 (28%)
Modified BTS alone
21 (16%)
No prior operation
11 (8%)
Main pulmonary artery banding
6 (5%)
Modified BTS and TAPVC repair
4 (3%)
Hybrid palliation
2 (2%)
Central shunt
2 (2%)
RVOT patch and modified BTS
1 (1%)
AV indicates atrioventricular; RVPA, right ventricle to pulmonary artery; BTS,
Blalock-Taussig shunt; TAPVC, total anomalous pulmonary venous connection;
and RVOT, right ventricular outflow tract.
*Age, height, and weight are from time of echo.
Stern et al
Table 2. Factors Associated With a Complete
Vascular Echocardiogram
Complete
(n⫽56, 43%)
Incomplete
(n⫽74, 57%)
Sex
0.16
Male
35 (62%)
36 (49%)
Female
21 (38%)
38 (51%)
Age, mo
Height, cm, n⫽34, 50
Weight, kg
BSA, m2
Time echo to
catheterization, d
5 (1–32)
5 (2–61)
0.80
61 (54–72)
62 (47–82)
0.43
identified by echo, 2 of which were accurately categorized.
None of the 10 cases of mild RPA stenosis were identified by
echo (5 not visualized, 5 no stenosis). For the 7 cases of
moderate RPA stenosis, only 2 were diagnosed by echo (3 not
visualized, 1 no stenosis, 1 mild stenosis). All 3 cases of RPA
stenosis identified by echo were confirmed by catheterization.
In the 81 studies in which the RPA was adequately imaged,
agreement with catheterization was 91% (n⫽74), though 6
stenoses were missed.
5.7 (2.4–12.0)
5.6 (2.4–19.7)
0.52
Echo Assessment of Central PA
0.31 (0.18–0.73)
0.51
1 (⫺57 to 51)
0 (⫺29 to 55)
0.69
The central PA was imaged by echo in 105 (81%) patients. Of
the 105 visualized by echo, agreement with catheterization
for assessment of degree of central PA distortion was 66%
(n⫽69). By catheterization, 82 central PAs were undistorted;
by echo, 59 of these were similarly categorized as undistorted, 4 were categorized as minor distortion, and 19 could
not be visualized. Minor central PA distortions were diagnosed in 35 by catheterization; by echo, 8 were successfully
identified, 19 were described as no distortion, 1 as major
distortion, and 7 could not be visualized. Of the 13 major
distortions diagnosed by catheterization, 2 were successfully
identified by echo, 9 were labeled as no distortion, and 2 as
mild distortion.
There were 45 discrete stenoses of the central PA diagnosed by catheterization. By echo, 37 central PAs were
identified as stenotic; however, only 20 of these stenoses
were corroborated on catheterization, and the other 17 were
mislabeled by echo as stenotic. Overall, the agreement of
echo with catheterization in appropriately identifying presence or absence of central PA stenosis was 59% (n⫽77).
0.28
Downloaded from http://circimaging.ahajournals.org/ by guest on May 11, 2017
Before
catheterization
41 (73%)
61 (82%)
After
catheterization
15 (27%)
13 (18%)
34 (61%)
37 (50%)
Complex 2V
3 (5%)
11 (15%)
DILV
6 (11%)
5 (7%)
Anatomic diagnosis
0.60
Single RV
5 (9%)
6 (8%)
Tricuspid atresia
3 (5%)
6 (8%)
Unbalanced CAVC
3 (5%)
5 (7%)
PA/IVS
2 (4%)
4 (5%)
34 (61%)
49 (66%)
Systemic shunt
9 (16%)
19 (26%)
None
8 (14%)
3 (4%)
Other*
5 (9%)
3 (4%)
Prior operation
Stage I
501
0.32 (0.18–0.52)
Echo timing
HLHS
P Value
Echo Before Glenn Operation
0.09
Sedation status
Echo Assessment of the Aortic Arch
0.72
Sedated study
30 (54%)
43 (58%)
Unsedated study
26 (46%)
31 (42%)
BSA indicates body surface area; HLHS, hypoplastic left heart syndrome; 2V,
2-ventricular anatomy; DILV, double-inlet left ventricle; RV, right ventricle;
CAVC, common atrioventricular canal; and PA/IVS, pulmonary atresia with
intact ventricular septum.
Data are depicted as No. (%) or median (range).
*Includes main pulmonary artery banding (n⫽6) and bilateral pulmonary
artery banding and stenting of the ductus arteriosus as part of hybrid stage I
palliation for hypoplastic left heart syndrome (n⫽2).
diagnosed by echo (8 not visualized, 1 no stenosis, 1
moderate stenosis). For moderate LPA stenosis, 2 of 8 cases
were diagnosed by echo (2 not visualized, 4 no stenosis). The
single case of LPA atresia was not identified by echo. Of the
6 cases of LPA stenosis identified by echo, 5 were confirmed
by catheterization and 1 was not. In the 83 studies in which
the LPA was adequately imaged, agreement with catheterization for presence or absence of stenosis was 92% (n⫽76),
though 5 stenoses were missed.
The RPA was imaged by echo in 81 (62%) patients. RPA
stenosis was present by catheterization in 17 patients (13%):
mild stenosis in 10 (8%), moderate stenosis in 7 (5%), no
severe stenosis or atresia. Of 17 RPA stenoses, only 3 were
The distal aortic arch was imaged adequately by echo in 104
(80%) studies. The proximal descending aorta was visualized
on 102 (78%) studies.
The distal aortic arch was adequately imaged by catheterization in 128 patients; 27 aortic arch obstructions were
diagnosed. Of these, 21 were identified by echo and 6 were
not (3 not visualized, 3 labeled as unobstructed.) Of the 103
studies in which the aortic arch was imaged by both echo and
catheterization, agreement with regard to presence or absence
of obstruction was 83% (85 of 103). Echo tended to overestimate arch obstruction, with 15 unobstructed arches labeled
as obstructed.
Vascular Measurements
Table 3 and Table 4 summarize the echo and catheterization
measurements of the PAs and aortic arch. Compared with
catheterization, the diameters of the PAs measured by echo
were generally smaller. The mean LPA measurement by echo
was 5.5 mm and by catheterization 6.2 mm (P⫽0.004)
(Figure 3). Similarly, the mean RPA measurement by echo
was 5.7 mm and by catheterization was 6.6 mm (P⫽0.001)
(Figure 4).
For the assessment of branch PA hypoplasia, there were 23
RPA measurements by echo that resulted in a z-score ⬍⫺2.
Only 3 of these measurements had corresponding catheterization measurements that also resulted in a z-score ⬍⫺2. For
502
Table 3.
Circ Cardiovasc Imaging
September 2011
Unobstructed Branch Pulmonary Artery Measurements
All studies
Sedated studies
Unsedated studies
Location of
Measurement
n
Echo, mm,
Mean⫾SD
Catheterization, mm,
Mean⫾SD
Mean Difference, mm,
Mean⫾SD
95% Confidence
Interval
P
Value
Pearson r*
RPA
67
5.7⫾1.8
6.6⫾2.4
⫺0.9⫾2.1
(⫺1.4, ⫺0.4)
0.001
0.56
LPA
68
5.5⫾1.9
6.2⫾2.4
⫺0.7⫾1.7
(⫺1.0, ⫺0.2)
0.004
0.72
CPA
89
6.2⫾3.2
6.5⫾3.4
⫺0.3⫾2.2
(⫺0.7, 0.2)
0.28
0.78
RPA
35
5.7⫾1.7
6.7⫾2.5
⫺1.0⫾2.1
(⫺1.7, ⫺0.3)
0.01
0.55
LPA
41
5.5⫾1.6
6.4⫾2.3
⫺0.9⫾1.9
(⫺1.5, ⫺0.3)
0.005
0.58
CPA
52
5.9⫾2.8
6.3⫾3.3
⫺0.4⫾2.2
(⫺1.0, 0.2)
0.24
0.76
RPA
32
5.7⫾2.0
6.5⫾2.4
⫺0.7⫾2.1
(⫺1.5, 0.03)
0.06
0.57
LPA
27
5.6⫾2.3
5.8⫾2.5
⫺0.2⫾1.1
(⫺0.6, 0.3)
0.41
0.90
CPA
37
6.6⫾3.8
6.7⫾3.6
⫺0.1⫾2.2
(⫺0.9, 0.6)
0.78
0.82
RPA indicates right pulmonary artery; LPA, left pulmonary artery; and CPA, central pulmonary artery.
*P values for Pearson correlation are all statistically significant (P⬍0.001).
Downloaded from http://circimaging.ahajournals.org/ by guest on May 11, 2017
the LPA, there were 19 measurements by echo that resulted in
z-scores ⬍⫺2; of these, 6 were confirmed by catheterization.
Agreement for central PA measurements was better. The
mean central PA measurement by echo was 6.2 mm and by
catheterization was 6.5 mm (P⫽0.28).
The size of the distal aortic arch was underestimated by
echo, with a mean of 6.1 mm versus 6.5 mm by catheterization (P⫽0.03, Figure 5). Echo and catheterization measurements of the proximal descending aorta were similar.
Sedation and Echo Timing
The use of sedation was not significantly associated with
increased frequency of complete echocardiograms. For the
LPA, unsedated echo measurements appeared to correlate
better with catheterization than sedated measurements. There
was no such effect noted for the RPA or central PA (Table 3).
Measurements of the distal aortic arch were more similar to
catheterization in echos performed under sedation; there was
no such significant difference for the proximal descending
aorta (Table 4).
Rates of visualization of vascular structures were not
significantly different in echos performed before versus after
catheterization. Rates of stenosis of the PAs, prevalence of
central PA distortion, and aortic arch obstruction by catheterization were also similar between the 2 groups (Table 5).
Interventions
Twenty-three patients with catheterization interventions had
precatheterization echos: 8 underwent balloon dilation of the
Table 4.
PAs and 17 underwent balloon dilation of the aortic arch.
Two patients had both the aortic arch and PAs dilated. Of the
8 patients with PA dilations, only 1 had an echo on which PA
stenosis was identified (3 not identified, 4 not visualized). Of
the 17 patients with arch dilations, 13 had arch obstruction
identified on echo (1 no obstruction, 3 not visualized).
Interventions to augment the PAs at time of BDG were
significantly associated with PA stenosis as well as central
PA distortion. Of those who had surgical PA augmentation,
87% had PA stenosis present on catheterization, versus 29%
in those without such interventions (P⬍0.001). Patients with
either mild or major central PA distortions were more likely
to undergo PA intervention (20 of 35 with mild distortion, 9
of 13 with major distortion; P⫽0.001).
Twelve patients had central PA distortion without central
or branch PA stenosis. Of these 12, 1 had an intervention on
the PAs at the time of BDG in the form of PA plasty without
patch material (rate of intervention on the PAs, 8%). This rate
of intervention on the PAs was significantly lower than for
patients with both central PA distortion and at least 1 PA
stenosis (P⬍0.001).
Of the 29 patients with branch PA stenosis not identified
on echo, all but 3 had surgical interventions to augment the
PAs at time of BDG. One of these 3 had balloon dilation of
the LPA at time of catheterization. Of the 6 patients with
aortic arch obstruction not identified on echo, 3 had no arch
intervention at time of BDG, all of whom underwent balloon
angioplasty of the coarctation at time of catheterization.
Distal Aortic Arch and Proximal Descending Aorta Measurements
All studies
Sedated studies
Unsedated studies
Location of
Measurement
n
Echo, mm,
Mean⫾SD
Catheterization,
mm, Mean⫾SD
Mean Difference,
mm, Mean⫾SD
95% Confidence
Interval
P
Value
Pearson r*
Distal arch
91
6.1⫾1.8
6.5⫾1.9
⫺0.4⫾1.5
(⫺0.7, ⫺0.03)
0.03
0.65
Proximal AoDt
89
7.6⫾1.4
8.0⫾2.0
⫺0.4⫾1.9
(⫺0.8, 0.02)
0.06
0.41
Distal arch
52
6.2⫾1.7
6.5⫾1.8
⫺0.2⫾1.6
(⫺0.7, 0.2)
0.29
0.56
Proximal AoDt
50
7.6⫾1.4
8.0⫾2.1
⫺0.5⫾2.1
(⫺1.0, 0.1)
0.13
0.38
Distal arch
39
6.0⫾2.1
6.5⫾2.0
⫺0.5⫾1.5
(⫺1.0, ⫺0.03)
0.04
0.75
Proximal AoDt
39
7.7⫾1.5
8.0⫾1.8
⫺0.3⫾1.7
(⫺0.8, 0.3)
0.31
0.45
AoDt indicates descending aorta.
*P values for Pearson correlation are all statistically significant (Pⱕ0.007).
Stern et al
Figure 3. Bland-Altman plot of right pulmonary artery measurements. Solid line represents the mean difference; dashed lines
represent the limits of agreement.
Downloaded from http://circimaging.ahajournals.org/ by guest on May 11, 2017
There were 17 patients in the cohort with no prior surgical
interventions on the branch PAs (no prior operation or main
PA banding only). Of these 17, there were 2 cases of RPA
stenosis and 2 cases of LPA stenosis diagnosed by catheterization on 4 separate patients. Three of these patients had a
main PA band as their prior operation, but 1 patient had no
previous operations. Two of these patients underwent patch
PA plasty at time of BDG (including the patient with no prior
operation), and a third had the Glenn anastomosis created in
such a way to relieve PA stenosis.
Discussion
In this large, retrospective study of BDG candidates with a
broad range of cardiac diagnoses and prior surgical interventions, echo showed generally suboptimal performance for the
evaluation of the relevant vascular anatomy. Echo was
frequently unable to adequately image the mediastinal branch
PAs and missed the majority of branch PA stenoses. When
visualized, there were statistically significant differences
between measurements for dimensions of the branch PAs,
with general underestimation of the diameter of both branch
PAs by echo. Although the differences in individual cases
were occasionally large as demonstrated by the BlandAltman figure, the average difference across the cohort was
generally ⬍1 mm. In general, the use of sedation did not
appear to improve the performance of echo for these factors.
The ability of echo to image the central PA was better than for
the branch PAs. Moreover, echo measurements of the central
PA were more similar to those obtained by catheterization.
Echo Before Glenn Operation
503
Figure 5. Bland-Altman plot of distal aortic arch measurements.
Solid line represents the mean difference; dashed lines represent the limits of agreement. Sedated and unsedated measurements are differentiated by icons.
However, agreement in identifying presence or absence of
stenosis was only 59%. Moreover, agreement in the categorization of the degree of distortion of the central PA was
generally poor, and echo missed the majority of cases of
major distortion. Distal aortic arch measurements by echo
tended to be underestimates as compared with catheterization,
but the use of sedation improved echo performance. Echo
successfully diagnosed 78% of aortic arch obstructions diagnosed by catheterization.
Previous studies investigating the accuracy of echo imaging of the PAs in congenital heart disease have had conflicting results and have been performed in a variety of different
lesions. Huhta et al11 compared echo with catheterization in
65 patients with pulmonary atresia with a ventricular septal
defect and found that the LPA was adequately imaged in only
Table 5. Echocardiogram Performance and
Catheterization Findings
Echo
All
(n⫽130)
Before
Catheterization
(n⫽102)
After
Catheterization
(n⫽28)
P Value
Visualization by
echo, n (%)
LPA
83 (64)
65 (64)
18 (64)
1
RPA
81 (62)
63 (62)
18 (64)
1
CPA
105 (81)
83 (81)
22 (79)
0.79
DAA
104 (80)
82 (80)
22 (79)
0.8
Proximal
descending
aorta
102 (78)
79 (77)
23 (82)
0.8
Findings at
catheterization,
n (%)
Figure 4. Bland-Altman plot of left pulmonary artery measurements. Solid line represents the mean difference; dashed lines
represent the limits of agreement.
LPA stenosis
21 (16)
17 (17)
4 (14)
RPA stenosis
17 (13)
15 (15)
2 (7)
1
0.36
CPA stenosis
45 (35)
37 (36)
8 (29)
0.51
CPA distortion
48 (37)
40 (39)
8 (29)
0.38
Aortic arch
obstruction
27 (21)
24 (24)
3 (11)
0.19
LPA indicates left pulmonary artery; RPA, right pulmonary artery; CPA,
central pulmonary artery; and DAA, distal aortic arch.
504
Circ Cardiovasc Imaging
September 2011
Downloaded from http://circimaging.ahajournals.org/ by guest on May 11, 2017
16 of 65 (25%), but the RPA was visualized in 55 of 65
(85%), and echo measurements correlated well with angiography. Gutgesell et al12 prospectively studied 20 cyanotic
infants with echo before catheterization and compared the
ability of echo to visualize and measure the central PA,
branch PAs and aortic arch. Vessel size was slightly underestimated though these structures were visualized readily.
Fogel et al13 compared echo and MRI with catheterization in
36 patients with single ventricle across different stages of
Fontan palliation, including 16 pre-BDG. They found more
significant variation in PA measurements with echo than with
MRI.13 Greenberg et al14 compared echo with MRI in 20
patients after repair for tetralogy of Fallot and found that echo
performed poorly, with the right and left branch PAs not
visualized in 8 and 10 children, respectively. Stenoses were
also missed in 13 branch PAs.
Even given this history of echo’s suboptimal performance,
the rate of incomplete echocardiograms in the current study
was higher than in previous studies. In prior reports from this
institution, higher rates of visualization of vascular structures
are described, ranging from 73% to 93%.9,10 However, the
previous studies involved smaller, less contemporaneous
cohorts. In addition, echocardiograms for prior studies were
likely less focused on details of the vascular anatomy, and
were interpreted by multiple different clinical staff. Secondary review of primary image data with regard to all vascular
structures was not necessarily performed. Furthermore, the
process of making independent blinded measurements of
vascular dimensions in the current report likely lowered the
threshold for nonvisualization of a structure, resulting in a
lower rate of complete echos.
Analyzing echos in which the structure of interest was
visualized, agreement between echo and catheterization on
PA stenosis was ⬎90% for both right and left PAs. However,
this number is misleading because 5 LPA and 6 RPA stenoses
were still missed. A false sense of reassurance can be created
if a complete echocardiogram is obtained and no PA stenoses
are identified.
The current study revealed equally poor performance for
both RPA and LPA visualization. Prior literature describes
better performance for echo imaging of the RPA as compared
with the LPA.11,15 One possible explanation for this discrepancy is the higher prevalence of systemic to PA shunts in the
current cohort. A Blalock-Taussig shunt or Sano conduit was
present in 84% of patients. Because these shunts are often
inserted onto the RPA, this could hinder visualization of this
structure.
Reliable assessment of PA distortion is difficult because of
its qualitative nature. PA distortion has been defined in the
literature as “distorted or markedly hypoplastic peripheral or
central pulmonary arteries”16 as well as the nonuniform
growth and development of the PAs.17 PA distortion has been
shown to increase the risk of death or takedown of Fontan
repair and is considered an independent risk factor for
mortality.16,18 We found an increased rate of surgical interventions on the PAs in patients with central PA distortion;
however, when patients with PA stenoses were eliminated
from this group, the rate of intervention fell to 1 of 12, or 8%.
It is possible that central PA distortion may represent a
surrogate for PA stenosis because patients with distortion and
no stenosis had a significantly decreased rate of surgical
intervention on the branch PAs.
The small number of patients without previous surgical
interventions on the branch PAs limits our conclusions from
these data, though they do suggest that suspicion for PA
stenoses should be high in all BDG candidates. In addition,
surgeons at our institution do not routinely dissect the branch
PAs out completely at time of BDG or use intraoperative
angiography. Therefore, thorough knowledge of branch PA
anatomy adds significantly to the surgical planning process.
A complete anatomic assessment of the branch PAs with a
modality sufficiently sensitive to detect stenoses should be
performed in all patients undergoing BDG, regardless of
underlying anatomy or prior interventions. Alternatively,
intraoperative angiography may be used at time of BDG to
assess the PAs.
The performance of echo for the evaluation of the aortic
arch in this setting, particularly in those with prior stage I
palliation with aortic reconstruction, has been well described.5,19 Our study confirmed echo as a reasonably good
screening tool for arch obstruction, diagnosing 21 of 27 cases.
However, it also predicted arch obstruction in many patients
that was not present at catheterization. Some of this probably
is due to the typical increase in flow velocity across the distal
reconstructed aortic arch, which can result in the false
impression of arch obstruction.
Echo has a major role in the pre-BDG evaluation of the
atrial septum, the degree and mechanism of valvular regurgitation, and the evaluation of ventricular function. All of
these factors are crucial to the short- and long-term outcomes
after BDG. However, this study demonstrates the limitations
of echo in its ability to visualize mediastinal vascular structures, particularly the branch PAs. There are no readily
identifiable methods available to improve the performance of
echo in this regard because the limitations probably are due
mostly to acoustic interference by lung tissue. It is clear that
echo alone cannot be relied on as the sole investigation before
BDG. We suggest using echo as part of a multimodality
imaging evaluation, with techniques such as cardiac magnetic
resonance, catheterization or intraoperative angiography used
to further delineate the relevant vascular anatomy.
Limitations
This study is prone to all the limitations inherent in retrospective investigations. Because the diagnostic studies were
not performed with the intention of being the sole preoperative imaging test, it is difficult to ascertain whether a
prospective study with this intent would yield substantially
different findings. However, at our institution, both echo and
catheterization are approached with the intent of evaluating
all relevant structures in detail, regardless of other planned
studies.
Additionally, surgical intervention at time of BDG was
used as a surrogate for clinical severity, though it is not
certain that many minor PA stenoses warrant intervention.
Conclusions
In a large cohort of patients presenting for BDG, echo
evaluation frequently failed to image the mediastinal branch
Stern et al
PAs, missed the majority of branch PA stenoses, and underestimated branch PA size. Sedation did not appear to improve
the performance of echo for the branch PAs. The aortic arch
was visualized more readily and the majority of aortic arch
obstructions were identified by echo; however, arch obstruction tended to be overestimated on echo. Echo cannot be
relied on as the sole investigation before BDG. Other diagnostic imaging tools such as cardiac magnetic resonance or
catheterization angiography appear to still be necessary for
complete visualization of the relevant vascular anatomy.
Disclosures
None.
References
Downloaded from http://circimaging.ahajournals.org/ by guest on May 11, 2017
1. Fontan F, Fernandez G, Costa F, Naftel DC, Tritto F, Blackstone EH,
Kirklin JW. The size of the pulmonary arteries and the results of the
Fontan operation. J Thorac Cardiovasc Surg. 1989;98:711–719.
2. Knott-Craig CJ, Danielson GK, Schaff HV, Puga FJ, Weaver AL,
Driscoll DJ. The modified Fontan operation: an analysis of risk factors for
early postoperative death or takedown in 702 consecutive patients from
one institution. J Thorac Cardiovasc Surg. 1995;109:1237–1243.
3. Choussat A, Fontan F, Besse P, Vallot F, Chauve A, Bricaud H. Selection
criteria for Fontan procedure. In: Anderson RH, Shinebourne EA, eds.
Pediatric Cardiology. Edinburgh: Churchill Livingstone; 1978:559 –566.
4. Jonas RA. Intermediate procedures after first-stage Norwood operation
facilitate subsequent repair. Ann Thorac Surg. 1991;52:696 –700.
5. Fraisse A, Colan SD, Jonas RA, Gauvreau K, Geva T. Accuracy of
echocardiography for detection of aortic arch obstruction after stage I
Norwood procedure. Am Heart J. 1998;135:230 –236.
6. Nakanishi T. Cardiac catheterization is necessary before bidirectional
Glenn and Fontan procedures in single ventricle physiology. Pediatr
Cardiol. 2005;26:159 –161.
7. McMahon CJ, Eidem BW, Bezold LI, Vargo T, Neish SR, Bricker JT,
Kovalchin J, El-Said H. Is cardiac catheterization a prerequisite in all
patients undergoing bidirectional cavopulmonary anastamosis? J Am Soc
Echocardiogr. 2003;16:1068 –1072.
8. Fogel MA. Is routine cardiac catheterization necessary in the management
of patients with single ventricles across staged Fontan reconstruction?
No! Pediatr Cardiol. 2005;26:154 –158.
Echo Before Glenn Operation
505
9. Brown DW, Gauvreau K, Moran AM, Jenkins KJ, Perry SB, del Nido PJ,
Colan SD. Clinical outcomes and utility of cardiac catheterization prior to
superior cavopulmonary anastomosis. J Thorac Cardiovasc Surg. 2003;
126:272–281.
10. Brown DW, Gauvreau K, Powell AJ, Lang P, Colan SD, Del Nido PJ,
Odegard KC, Geva T. Cardiac magnetic resonance versus routine cardiac
catheterization before bidirectional Glenn anastomosis in infants with
functional single ventricle: a prospective randomized trial. Circulation.
2007;116:2718 –2725.
11. Huhta JC, Piehler JM, Tajik AJ, Hagler DJ, Mair DD, Julsrud PR, Seward
JB. Two dimensional echocardiographic detection and measurement of
the right pulmonary artery in pulmonary atresia-ventricular septal defect:
angiographic and surgical correlation. Am J Cardiol. 1982;49:1235–1240.
12. Gutgesell HP, Huhta JC, Cohen MH, Latson LA. Two-dimensional echocardiographic assessment of pulmonary artery and aortic arch anatomy in
cyanotic infants. J Am Coll Cardiol. 1984;4:1242–1246.
13. Fogel MA, Donofrio MT, Ramaciotti C, Hubbard AM, Weinberg PM.
Magnetic resonance and echocardiographic imaging of pulmonary artery
size throughout stages of Fontan reconstruction. Circulation. 1994;90:
2927–2936.
14. Greenberg SB, Crisci KL, Koenig P, Robinson B, Anisman P, Russo P.
Magnetic resonance imaging compared with echocardiography in the
evaluation of pulmonary artery abnormalities in children with tetralogy of
Fallot following palliative and corrective surgery. Pediatr Radiol. 1997;
27:932–935.
15. Vick GW, Rokey R, Huhta JC, Mulvagh SL, Johnston DL. Nuclear
magnetic resonance imaging of the pulmonary arteries, subpulmonary
region, and aorticopulmonary shunts: a comparative study with twodimensional echocardiography and angiography. Am Heart J. 1990;119:
1103–1110.
16. Mayer JE, Bridges ND, Lock JE, Hanley FL, Jonas RA, Castaneda AR.
Factors associated with marked reduction in mortality for Fontan operations in patients with single ventricle. J Thorac Cardiovasc Surg. 1992;
103:444 – 452.
17. Jonas RA. Intermediate procedures after first-stage Norwood operation
facilitate subsequent repair. Ann Thorac Surg. 1991;52:696 –700.
18. Bridges ND, Jonas RA, Mayer JE, Flanagan MF, Keane JF, Castaneda
AR. Bidirectional cavopulmonary anastamosis as interim palliation for
high-risk Fontan candidates. Circulation. 1990;82(suppl IV):IV170 –IV-176.
19. Sekar P, Border WL, Kimball TR, Hirsch R, Manning PB, Khoury PR,
Beekman RH III. Aortic arch recoarctation after the Norwood stage I
palliation: the comparative accuracy of blood pressure cuff and echocardiographic Doppler gradients in detecting significant obstruction.
Congenit Heart Dis. 2009;4:440 – 447.
CLINICAL PERSPECTIVE
There has been increasing interest in noninvasive evaluation alone in patients with single-ventricle circulation before
undergoing bidirectional Glenn operation. We studied the ability of echocardiography to visualize and assess the relevant
vascular anatomy in this patient population. Echocardiography was found to perform poorly compared with catheter
angiography. The branch pulmonary arteries were successfully imaged by echocardiography in under two-thirds of
patients, and the majority of pulmonary artery stenoses found at catheterization were not visualized by echocardiography.
The aortic arch was imaged more readily by echocardiography, with the majority of arch obstructions identified. Sedation
did not appear to improve the performance of echocardiography for assessment of the pulmonary arteries. Given the clinical
importance of identification and treatment of obstructions to pulmonary blood flow in the single ventricle circulation, we
conclude that echocardiography alone before bidirectional Glenn operation is insufficient to image the relevant vascular
anatomy.
Echocardiographic Evaluation Before Bidirectional Glenn Operation in Functional
Single-Ventricle Heart Disease: Comparison to Catheter Angiography
Kenan W.D. Stern, Doff B. McElhinney, Kimberlee Gauvreau, Tal Geva and David W. Brown
Downloaded from http://circimaging.ahajournals.org/ by guest on May 11, 2017
Circ Cardiovasc Imaging. 2011;4:498-505; originally published online July 5, 2011;
doi: 10.1161/CIRCIMAGING.110.963280
Circulation: Cardiovascular Imaging is published by the American Heart Association, 7272 Greenville Avenue,
Dallas, TX 75231
Copyright © 2011 American Heart Association, Inc. All rights reserved.
Print ISSN: 1941-9651. Online ISSN: 1942-0080
The online version of this article, along with updated information and services, is located on the
World Wide Web at:
http://circimaging.ahajournals.org/content/4/5/498
Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published
in Circulation: Cardiovascular Imaging can be obtained via RightsLink, a service of the Copyright Clearance
Center, not the Editorial Office. Once the online version of the published article for which permission is being
requested is located, click Request Permissions in the middle column of the Web page under Services. Further
information about this process is available in the Permissions and Rights Question and Answer document.
Reprints: Information about reprints can be found online at:
http://www.lww.com/reprints
Subscriptions: Information about subscribing to Circulation: Cardiovascular Imaging is online at:
http://circimaging.ahajournals.org//subscriptions/