Accepted Manuscript A Pulsatile Hemodynamic Evaluation of the Commercially Available Bifurcated YGraft Fontan Modification and Comparison to the Lateral Tunnel and Extracardiac Conduits Phillip M. Trusty, BS, Maria Restrepo, PhD, Kirk R. Kanter, MD, Ajit P. Yoganathan, PhD, Mark A. Fogel, MD, Timothy C. Slesnick, MD PII:
S0022-5223(16)00418-9
DOI:
10.1016/j.jtcvs.2016.03.019
Reference:
YMTC 10407
To appear in:
The Journal of Thoracic and Cardiovascular Surgery
Received Date: 16 December 2015 Revised Date:
1 March 2016
Accepted Date: 5 March 2016
Please cite this article as: Trusty PM, Restrepo M, Kanter KR, Yoganathan AP, Fogel MA, Slesnick TC, A Pulsatile Hemodynamic Evaluation of the Commercially Available Bifurcated Y-Graft Fontan Modification and Comparison to the Lateral Tunnel and Extracardiac Conduits, The Journal of Thoracic and Cardiovascular Surgery (2016), doi: 10.1016/j.jtcvs.2016.03.019. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
A Pulsatile Hemodynamic Evaluation of the Commercially Available Bifurcated Y-Graft Fontan Modification and Comparison to the Lateral Tunnel and Extracardiac Conduits
1 2 3 4 5 6
35 36 37 38 39 40 41
RI PT
M AN U
SC
Wallace H. Coulter Department of Biomedical Engineering, Georgia Institute of Technology and Emory University, Atlanta, GA 2 Division of Cardiothoracic Surgery, Children’s Healthcare of Atlanta, Emory University School of Medicine, Atlanta, GA, USA. 3 Division of Cardiology, Children’s Hospital of Philadelphia, Philadelphia, PA, USA 4 Department of Pediatrics, Division of Cardiology, Children’s Healthcare of Atlanta, Emory University School of Medicine, Atlanta, GA, USA
TE D
Abbreviations and Acronyms: BSA (body surface area), CFD (computational fluid dynamics), CMR (cardiac magnetic resonance), ECC (extra cardiac conduit), HFD (hepatic flow distribution), iPL (indexed power loss), IVC (inferior vena cava), LPA (left pulmonary artery), LT (lateral tunnel), PA (pulmonary artery), PAVMs (pulmonary arteriovenous malformations), PFD (pulmonary flow distribution), PI (pulsatility index), PL (power loss), RPA (right pulmonary artery), SVC (superior vena cava), TCPC (total cavopulmonary connection)
Funding/Disclosures: P.M. Trusty: None. M. Restrepo: None. M.A. Fogel: Research Grant; Modest; NIH R01. Consultant/Advisory Board; Modest; Edwards Life Sciences - MRI Core Lab. Other; Modest; AMAG FACT trial site, Cooley's Anemia Foundation MRI Core Lab. K. Kanter: None. A.P. Yoganathan:None. T.C. Slesnick: None. Conflict of Interest: None of the authors have potential conflicts of interest in relation to the presented work.
EP
26 27 28 29 30 31 32 33 34
1
AC C
7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Phillip M. Trusty, BS1; Maria Restrepo, PhD1; Kirk R. Kanter, MD2; Ajit P. Yoganathan, PhD1; Mark A. Fogel, MD3; Timothy C. Slesnick, MD4
Corresponding Author: Ajit P. Yoganathan
[email protected] 404-502-7869
Article Word Count: 3343 1
ACCEPTED MANUSCRIPT
Abstract
43
Word count: 225
44
Objective: Fontan completion, resulting in a total cavopulmonary connection (TCPC), is
45
accomplished using either a lateral tunnel (LT), extracardiac conduit (ECC), or bifurcated Y-graft.
46
The use of Y-grafts is hypothesized to provide symmetric hepatic blood flow distribution to the
47
lungs, a factor related to pulmonary arteriovenous malformations. The present study evaluates
48
the hemodynamic performance of the largest commercially available Y-graft cohort to date,
49
highlights 6 representative cases and compares commercially available Y-graft performance
50
with LT/ECC connections.
51
Methods: Thirty commercially available Y-graft and 30 LT/ECC patients were analyzed. TCPC
52
anatomies and flow waveforms were reconstructed using cardiac magnetic resonance (CMR)
53
images and phase-contrast CMR. Computational fluid dynamic simulations were performed to
54
quantify TCPC power loss, resistance, and hepatic flow distribution (HFD). Comparisons
55
between graft types were investigated.
56
Results: TCPC resistance was significantly higher for Y-grafts. HFD was similar overall, but shows
57
discrepancies at extreme values with more unbalanced flow in the Y-graft cohort. Power loss
58
was more sensitive to LPA stenosis in the Y-graft cohort. Prediction of Y-graft HFD is multi-
59
factorial.
60
Conclusions: Commercially available Y-grafts do not inherently provide more balanced HFD than
61
LT/ECC connections, which are more energetically favorable and less sensitive to PA stenosis.
62
Graft type should be considered on an individual basis as hemodynamic performance is based
AC C
EP
TE D
M AN U
SC
RI PT
42
2
ACCEPTED MANUSCRIPT
63
on a combination of factors, including pulmonary flow distribution, PA stenosis and superior
64
vena cava positioning.
RI PT
65
Central Message:
67
Commercially available Y-grafts do not inherently provide more balanced HFD or favorable
68
energetic performance than traditional Fontans.
SC
66
AC C
EP
TE D
M AN U
69
3
ACCEPTED MANUSCRIPT
Perspective Statement:
71
Despite intuition, commercially available Y-grafts do not inherently provide more balanced
72
hepatic flow distribution than traditional Fontan connections. Several Y-grafts were seen to
73
direct nearly all hepatic flow to a single lung. Graft type should be considered on an individual
74
basis as hemodynamic performance is governed by several factors including pulmonary flow
75
distribution and PA stenosis.
SC
RI PT
70
76
AC C
EP
TE D
M AN U
77
4
ACCEPTED MANUSCRIPT
78
Introduction The Fontan procedure is the final operation in a series of palliative surgical procedures
80
for patients with single ventricle congenital heart defects, in which the inferior vena cava (IVC)
81
is connected to the pulmonary arteries (PA) by way of an extra-cardiac conduit or intra-cardiac
82
tunnel. This operation results in the total cavopulmonary connection (TCPC) which bypasses the
83
ventricle with passive venous return into the PAs.1 With advances in medical imaging, clinicians
84
can better determine the appropriate course of action for individual patients prior to surgery.
85
Previous studies focused on surgical planning for the Fontan operation in an attempt to
86
maximize energy efficiency, obtain equal hepatic flow distribution (HFD) and potentially
87
improve the patient’s long term quality of life.2–4 Further efforts characterized the effect of
88
baffle diameter, vessel offset, connection angle and surgical approach on the power loss and
89
HFD of the TCPC.5–10
SC
M AN U
TE D
90
RI PT
79
Currently, surgeons complete the Fontan connection using either a lateral tunnel (LT) baffle, extracardiac conduit (ECC) or Y-graft. The ECC is typically positioned externally around
92
the margin of the heart while the LT is positioned through the atrium. LT/ECC connections are
93
usually anastomosed to the mid-PA, inferior to the superior vena cava (SVC) anastomosis from
94
the prior bidirectional Glenn. Alternatively, the commercially available bifurcated Y-graft
95
modification splits inferior systemic venous return through two branches toward the left and
96
right pulmonary arteries. Y-Grafts are utilized in an effort to obtain balanced HFD to the left and
97
right lungs, a factor in preventing pulmonary arteriovenous malformations (PAVMs), a potential
98
Fontan patient morbidity.11,12 The TCPCs from two patients employing each option are shown in
99
Figure 1.
AC C
EP
91
5
ACCEPTED MANUSCRIPT
100 101
For the inherently inefficient circulation in Fontan patients, all attempts to minimize the ventricle’s workload are clinically valuable. Surgeons can repair branch pulmonary artery
103
stenosis, determine energetically favorable baffle placement, and choose which connection
104
type to use. However, surgeons may not have complete freedom in these choices as they are
105
limited by the specific patient’s anatomy and condition.
SC
106
RI PT
102
The potential energetic and HFD differences associated with graft choice have not been fully analyzed in large cohorts under pulsatile simulation conditions. Previous studies showed
108
exploratory and preliminary results discussing the effectiveness of Y-graft use, but included very
109
small sample sizes.13–18 Some of these studies employed time-averaged boundary conditions
110
which can cause non-negligible differences in TCPC hemodynamics.9,19 The present study
111
compares hemodynamic performance between LT/ECC and commercially available Y-graft
112
TCPCs with the largest patient cohort to date while employing patient-specific transient flow
113
boundary conditions in order to provide insights into graft choice for Fontan completion.
114
Methods
116
Patient Cohort
117
AC C
115
EP
TE D
M AN U
107
A total of 60 patients were used in this study. All LT/ECC patient data (n=30) were
118
received from Children’s Healthcare of Philadelphia and all Y-Graft patient data (n=30) were
119
received from Children’s Healthcare of Atlanta. Thirty consecutive Y-grafts performed by Dr.
120
Kirk Kanter were used; LT/ECC patients were chosen to match age and BSA between groups. All
121
Y-graft Fontan connections used commercially available Y-grafts. Clinical data including age,
6
ACCEPTED MANUSCRIPT
122
gender, and BSA were obtained for each patient. The Institutional Review Boards of the
123
institutions involved approved this study.
125
RI PT
124
Anatomy and Velocity Reconstruction
Patient-specific anatomies were reconstructed from axial CMR images using methods
127
previously developed.20,21 Geomagic Studio® (Geomagic Inc., Research Triangle Park, NC) was
128
then used to fit a surface around the reconstructed point-cloud and export the surface for mesh
129
generation. Patient-specific blood flow waveforms were calculated from phase contrast MRI
130
images for all vessels of interest using previously validated methods.22,23
131
133
Computational Fluid Dynamics
All geometries were imported into GAMBIT (ANSYS Inc., Lebanon, NH) for mesh
TE D
132
M AN U
SC
126
generation. To provide computational stability, flow extensions of 10mm and 50mm were
135
added to all inlets and outlets, respectively. All computational fluid dynamics (CFD) simulations
136
were completed using a previously developed and validated in-house immersed boundary
137
solver.4,5,14,24 The unsteady flow waveforms extracted from PC-MRI were imposed as boundary
138
conditions for each inlet and outlet. A total of 6 cardiac cycles were simulated for each patient
139
to overcome transition effects and achieve periodic stability; the final cycle was used for
140
hemodynamic analysis.
AC C
EP
134
141 142 143
7
ACCEPTED MANUSCRIPT
145 146
Data Analysis Calculations were performed to compute power loss (PL), indexed power loss (iPL), TCPC resistance, pulsatility index (PI) and percent stenosis. Power loss was calculated as follows:
RI PT
144
147
where p is the static pressure, ρ is the blood density, A is the area of the inlet/outlet and v the
149
velocity. Power loss was calculated for 200 points throughout the cardiac cycle and then
150
averaged to obtain an overall PL value. Power loss was indexed (iPL) as (
151
resistance was calculated
152
surface area (m2), ∆ is the pressure drop and ρ is the blood density (kg·m-3).
153
Pulsatility index was defined as follows:
/
) and a TCPC
M AN U
∆
SC
148
where Qs is the systemic venous flow (L·s-1), BSA is the body
2 ∗ !"
TE D
where Qmax, Qmin, and Qavg are the maximum, minimum and average flow rates during the
155
cardiac cycle, respectively, as determined from the phase contrast CMR data. Percent stenosis
156
was calculated for the pulmonary arteries using the following equation:
EP
154
157
AC C
#$%&'()( % 100 -
., 0 .,1
2 .!", 0 .!",1
where ., and .,1 are the minimum cross sectional areas of the LPA and RPA, and
158
.!", and .!",1 are the average cross sectional areas of the LPA and RPA. A combined
159
outlet stenosis value was used in place of unilateral PA stenosis measurements to give a more
160
accurate representation of the total outlet obstruction to flow and allow a better comparison
161
between patients. Additionally, IVCs and SVCs showed negligible percent stenosis and therefore 8
ACCEPTED MANUSCRIPT
162
were not included in this calculation. All vessel areas for this metric were calculated using
163
Vascular Modeling Toolkit version 1.0.1 (Orobix, Bergamo, Italy). Hepatic flow distribution (HFD) was evaluated using particle tracking. Particles were
RI PT
164
seeded at the base of the LT/ECC or Y-graft at 200 time points throughout the cardiac cycle. An
166
additional four cardiac cycles (with no particle seeding) were simulated to ensure that the
167
residence time of the particles would not exceed the simulation time. Particles exiting at each
168
outlet were counted and HFD was defined as the ratio of particles leaving the LPA: 67$)89%(
67$)89%( 0 67$)89%(1
M AN U
345
SC
165
HFD will be discussed in terms of “HFD deviation from 50%”, and an even (50/50) split of
170
hepatic flow to the LPA and RPA is assumed as ideal. This convention gives a “0” HFD deviation
171
for balanced HFD cases, and an HFD deviation of 50 for extreme cases (all hepatic flow to either
172
the LPA or RPA). Particle residence time was calculated by recording the amount of time
173
required for each particle to exit the TCPC domain. An overall pulmonary flow distribution (PFD)
174
was calculated as the ratio of LPA flow to total pulmonary flow:
TE D
169
176 177
:;< =<:>?
AC C
175
EP
45
Statistics
The Anderson-Darling test was used to determine normality for each parameter within
178
both patient groups. Depending on normality, either a Wilcoxon rank sum test (non-normal
179
distribution) or a two-sample t-test (normal distribution) was used to test for equal medians
9
ACCEPTED MANUSCRIPT
180
between patient groups using SPSS (IBM Corp., Version 23, Armonk, NY). An asterisk (*) is used
181
to signify statistical significance using α=0.05. Values are shown as mean± standard deviation.
183
RI PT
182
Results
184
The commercially available Y-graft evaluation includes an investigation of several
186
hemodynamic metrics within the Y-graft cohort, then an examination of both ideally and poorly
187
performing Y-grafts (in terms of HFD), and finally, a comparison between Y-graft and LT/ECC
188
performance.
189
191
Y-Graft Hemodynamic Performance
Figure 2 shows several clinically important hemodynamic metrics for the Y-graft cohort.
TE D
190
M AN U
SC
185
The average PFD and HFD deviations from an ideal 50/50 split were 13.4±10.6 and 25.9±15.1,
193
respectively (Fig. 2a). Fig. 2b illustrates the effect of LPA stenosis on indexed power loss. A
194
moderate positive correlation (r=0.67, p<0.01) was found between LPA stenosis and indexed
195
power loss. The distributions of TCPC resistance and HFD deviation are shown in Fig. 2c and 2d.
196
The distribution of TCPC resistance is skewed to the right with approximately 53% and 70% of Y-
197
graft patients having resistances less than 1 and 2 Wood Units respectively. HFD deviation was
198
relatively evenly distributed (Fig. 2d). Only 13% of Y-graft patients have HFD deviations less
199
than 10 (ideal) and 24% of Y-graft patients have deviations greater than 40 (poor).
AC C
EP
192
200 201
Representative Y-Graft Cases
10
ACCEPTED MANUSCRIPT
202
Figure 3 shows streamlines for six representative cases from the 30 patient Y-graft cohort. Patients Y1, Y2, andY3 show balanced HFD (low HFD deviation) while the final three (Y4,
204
Y5, andY6) are unbalanced. Y1 represents an “ideal” case with balanced PFD, normal SVC
205
positioning and no mid-PA stenosis. Y2 represents uneven PFD with mid-PA stenosis on the low
206
PFD side. Y3 has bilateral SVCs, balanced PFD and mid-PA stenosis. Y4 has balanced PFD with
207
the SVC positioned superior to a y-branch anastomosis. Y5 has unbalanced PFD with mid-PA
208
stenosis on the high PFD side and Y6 represents balanced PFD with y-branch stenosis. HFD to
209
the LPA is noted for each model in Figure 3.
M AN U
210 211 212
SC
RI PT
203
Y-graft vs LT/ECC
To further evaluate the commercially available Y-graft, a direct hemodynamic comparison was performed between the Y-graft and LT/ECC cohorts. Clinical data are shown in
214
Table 1 for each graft type. Patients were selected to best match age and BSA between groups.
215
Y-graft resistance was significantly higher (p=0.03) than LT/ECC resistance, with averages of
216
1.51±1.52 and 0.56±0.54 WU, respectively (Fig. 4a). Overall, HFD deviation was not significantly
217
different between the groups with an average of 21±13 for the LT/ECC and 26±15 for Y-grafts
218
(Fig. 4b). Figure 4c shows no significant differences in particle residence time.
EP
AC C
219
TE D
213
Figure 4d and 4e compare the distribution of TCPC resistance and HFD deviation
220
between graft types. Almost 77% of LT/ECC patients have a resistance less than 1 WU,
221
compared to 53% of the Y-grafts. Additionally, 30% of commercially available Y-graft patients
222
exhibit a resistance greater than 2 WU, while only 3% of the LT/ECC cohort have values this high
223
(Fig. 4d). Focusing on the two extremes of HFD deviation, 30% of LT/ECC patients had ideal HFD
11
ACCEPTED MANUSCRIPT
deviations (<10) compared to only 10% of Y-graft patients. At the other extreme, more than
225
twice as many Y-graft patients (23%) have HFD deviations greater than 40 (poor) than LT/ECCs
226
(10%).
227
Figure 4f compares the sensitivity of iPL to LPA stenosis for both graft types. Commercially
228
available Y-grafts were found to be more sensitive to LPA stenosis in terms of iPL with a
229
significant difference in the two correlations (Fisher’s exact test, one-tailed, z=-1.76, p=0.04).
SC
RI PT
224
230
232 233
M AN U
231
Discussion
Capitalizing on the relatively large sample size in this study, the following section evaluates commercially available Y-graft performance and generalizes differences between
235
graft types. Both clinical and physiological implications are discussed. Six Y-grafts that illustrate
236
common anatomical parameter combinations will be emphasized and study limitations will be
237
considered.
240
EP
239
Y-Graft Hemodynamic Performance
As expected, PFD deviation is relatively low and on average there is about a 25%
AC C
238
TE D
234
241
difference in overall flow rates to the LPA and RPA. This flow split is determined partially by
242
branch pulmonary artery stenosis.25 HFD deviation was higher and more variable than
243
expected and will be discussed in the following section. The effect of LPA stenosis on iPL is also
244
an expected result. Blood accelerates through the stenosis region and then decelerates as the
245
vessel returns to its original size according to the conservation of momentum principle. These
12
ACCEPTED MANUSCRIPT
changes in velocity result in an increased pressure drop across the TCPC (lower efficiency)
247
which leads to an increased iPL. The LPA is a common site for stenosis and Fig. 2b shows that
248
the most severe LPA stenosis anatomies have the highest iPL.10
249
RI PT
246
The distribution of TCPC resistance raises concern for a number of Y-graft patients. In this study, Y-graft TCPCs are seen to have resistances as high as 4.7 Wood units, and this
251
increased level of vascular resistance could lead to venous hypertension/congestion. Liver
252
failure and protein-losing enteropathy are thought to be direct results of increased Fontan
253
venous pressures.26–28 The relatively even distribution of HFD deviation suggests that the Y-
254
graft is not a “one size fits all” solution to accomplish evenly split hepatic flow to the lungs.
255
While 13% of Y-graft patients had very balanced HFD, nearly a quarter of patients had >90%
256
hepatic flow directed to either the left or right lung. The effectiveness of Y-graft use seems to
257
be multi-factorial and extremely patient specific, suggesting the need for pre-operative,
258
individual surgical planning via computational modeling.
261
M AN U
TE D
260
Representative Y-Graft Cases
EP
259
SC
250
The ability of a Y-graft to evenly split hepatic flow is highly dependent on PFD, the position of branch PA stenosis and SVC flow collision, among other factors. A difference in PFD
263
can cause two Y-grafts with very similar anatomies to have drastically different HFDs. This can
264
be seen in Fig. 3, where Y2 and Y5 have PFDs of 24% and 65%, respectively. Despite similar
265
anatomies and Y-grafts, the HFD for these cases are considerably different (48% and 81% to the
266
LPA, respectively). A comparison between Y1 and Y4 illustrates the effect of SVC flow collision
267
on HFD. Both cases have balanced PFD and similar sized PAs, but Y1 has a medially positioned
AC C
262
13
ACCEPTED MANUSCRIPT
SVC, while Y4’s SVC is positioned more superior to the left Y-branch. The SVC flow collision in Y4
269
forces some SVC flow into the left Y-branch which causes hepatic flow to preferentially course
270
through the right Y-branch. The result is a very unbalanced HFD for Y4 (12% to the LPA) and
271
relatively balanced HFD for Y1 (65% to the LPA).
272
RI PT
268
Stenosis is commonly seen in Fontan anatomies;10 Y2, Y3, Y5 and Y6 all have severe stenosis in the TCPC. For single SVC anatomies (Y2, Y5), mid-PA stenosis influences SVC flow
274
which will preferentially follow the path of least resistance (in these cases towards the RPA).
275
For an unbalanced PFD case, if the mid-PA stenosis is located on the same side as the PA
276
receiving the majority of pulmonary flow, then the stenosis blocks most SVC flow and this PA
277
receives the majority of its flow from the hepatics and will therefore have an unbalanced HFD
278
(Y5). However, the opposite is true if the stenosis is located on the low PFD side; in this case
279
hepatic flow is able to move towards both PAs resulting in more balanced HFD (Y2). For
280
bilateral SVC anatomies (Y3), mid-PA stenosis can “separate” the two SVC flows and HFD will be
281
highly dependent on the individual SVC and overall PA flows. If SVC flows are similar and PFD is
282
balanced, a balanced HFD is expected. Y6 shows a case with severe stenosis located in the left
283
Y-branch, and as a result hepatic flow is very limited through the stenosed branch. With the full
284
range of flow rates, flow ratios, pulsatilities, unique geometries and stenosis locations, the
285
ability of a specific Y-graft to achieve balanced HFD may not be intuitive. Again, Y-graft
286
performance needs to be evaluated on an individual basis and all of these factors must be
287
considered.
AC C
EP
TE D
M AN U
SC
273
288 289
Y-graft vs LT/ECC
14
ACCEPTED MANUSCRIPT
290
Average TCPC resistance was approximately 3x higher for Y-grafts than LT/ECCs. The most apparent explanation stems from the geometry of the Y-graft itself. With commercially
292
available Y-grafts, the diameter of the base of the Y-graft is evenly split between the two
293
branches. For example, a 20mm (diameter) Y-graft will have two 10mm (diameter) branches. A
294
basic calculation reveals that preserving diameter does not preserve area; this design results in
295
a 50% reduction in cross-sectional area as blood moves into the Y-graft branches. In essence, Y-
296
grafts have a built-in 50% long segment “stenosis”. Additionally, resistance is inversely
297
proportional to the fourth power of the radius (Poiseuille’s law). Therefore, blood traveling
298
through 2 Y-branches with diameters half the size of the LT/ECC would analytically yield an 8x
299
increase in resistance. According to the results presented here, the reduction in diameter and
300
cross sectional area inherent to commercially available Y-grafts has a greater negative impact
301
on TCPC resistance than the venous flow collisions inherent to LT/ECCs. Therefore, “custom” Y-
302
grafts which preserve cross sectional area as shown by Yang et al. may be the more promising
303
option for successful Y-graft use, and may potentially provide an advantage over the LT/ECC
304
Fontan pathway.15,18
SC
M AN U
TE D
EP
305
RI PT
291
Overall, HFD deviation was not significantly different between graft types. Both graft types were highly variable. The most important observation is the discrepancy in the number of
307
patients at both the lowest (ideal) and highest (poor) HFD deviation (Fig. 4e). In this study, Y-
308
grafts have almost twice as many poor differential cases as ideal cases, while LT/ECCs show the
309
opposite trend. These results highlight the need for pre-operative planning with Y-graft use as
310
well as the design of individualized Y-grafts which are tailored for a specific patient. Potential
AC C
306
15
ACCEPTED MANUSCRIPT
311
modifications such as the angle of origin, branch size and overall shape could lead to better HFD
312
performance on an individual basis. Particle residence time (PRT) was investigated for its potential impact on flow-induced
314
thrombus formation, as well as to provide quantitative information about flow stagnation.29,30
315
No differences were seen for minimum, mean or maximum residence times. As seen in Fig. 4f,
316
Y-graft iPL was more sensitive to LPA stenosis than LT/ECC iPL. PA stenosis may have a greater
317
influence on the reduced flow through the closest Y-branch than it does on the total hepatic
318
flow traveling through a LT/ECC connection. This difference would lead to a greater effect of
319
stenosis on iPL for Y-grafts than LT/ECCs.
M AN U
SC
RI PT
313
Several additional comparisons between graft types are important to note. First, PFD
321
was identical between graft types at 46% to the left lung, indicating that graft type does not
322
affect PFD. Additionally, there was significantly more PA stenosis (p=0.02) in the LT/ECC cohort
323
as seen in Table 1. Even with significantly more PA stenosis, which is known to increase
324
resistance, the LT/ECC cohort still had lower TCPC resistance than the Y-grafts. This strengthens
325
the finding that commercially available Y-grafts inherently create higher TCPC resistance.
326
Finally, the LT/ECC group includes both intra-atrial and extracardiac connections, while Y-grafts
327
are always extracardiac. In this study, 10 of the LT/ECCs were intra-atrial. To ensure that these
328
results were a consequence of graft type and not the intra-atrial vs. extracardiac variation, a
329
sub-comparison was performed between these two groups. No statistically significant
330
differences were found in this sub-comparison for iPL, PFD, or HFD, and therefore the
331
comparative results presented here are due to graft choice and not its location.
AC C
EP
TE D
320
332
16
ACCEPTED MANUSCRIPT
Limitations
334
The period of time between the Fontan operation and the MRI scan was not identical for all
335
patients included in this study. Y-graft patients were scanned within one month of the surgery,
336
while LT/ECC patients were scanned around 6 months after the operation. This difference in
337
recovery time may lead to variations in the amount of physiological adaptation which could
338
affect flow rates (as seen in the difference between systemic venous flows) and pulmonary
339
vascular resistance and therefore TCPC resistance and HFD. Additionally, the spatial velocity
340
profiles used at the inlets of the CFD domain are assumed to be parabolic, which are not
341
identical to the spatial velocity profiles obtained from the MRI velocity segmentations.
342
However, this assumption is expected to have only minimal effects on the bulk hemodynamic
343
metrics. From a modeling perspective, this study focused only on the TCPC geometry and did
344
not include interactions with the rest of the cardiovascular or pulmonary circulation. This
345
localized TCPC modeling may omit significant global influences on the system. Finally, due to
346
limited data, the hepatic veins were not included in this modeling. The inclusion of the hepatic
347
veins and their respective flow rates could potentially influence the mixing of hepatic factor
348
with the IVC flow and therefore alter the presented HFD results. This study also assumes rigid
349
walls in the CFD simulations.
351
SC
M AN U
TE D
EP
AC C
350
RI PT
333
Conclusion
352 353 354
In this study, commercially available Y-grafts were found to have highly variable HFD due to a combination of factors including pulmonary flow distribution to the left and right
17
ACCEPTED MANUSCRIPT
lungs, presence and location of stenosis, and SVC positioning. HFD distributions suggest that
356
LT/ECCs are more likely to avoid extremely unbalanced hepatic flow to the lungs. Y-graft HFD
357
performance is multi-factorial, which requires patient specific computational modeling. Y-grafts
358
were found to be more sensitive in terms of indexed power loss to LPA stenosis. Overall, Y-
359
grafts show significantly higher resistance than LT/ECCs due in part to the intrinsic graft
360
geometry, which calls for a better Y-graft design.
SC
RI PT
355
361
AC C
EP
TE D
M AN U
362
18
ACCEPTED MANUSCRIPT
363
References
364
1.
doi:10.1136/thx.26.3.240.
366 367
Fontan F, Baudet E. Surgical repair of tricuspid atresia. Thorax. 1971;26(3):240-248.
RI PT
365
2.
Pekkan K, Whited B, Kanter K, et al. Patient-specific surgical planning and hemodynamic computational fluid dynamics optimization through free-form haptic anatomy editing
369
tool (SURGEM). Med Biol Eng Comput. 2008;46(11):1139-1152. doi:10.1007/s11517-008-
370
0377-0. 3.
M AN U
371
SC
368
Sundareswaran KS, de Zélicourt D, Sharma S, et al. Correction of Pulmonary
372
Arteriovenous Malformation Using Image-Based Surgical Planning. JACC Cardiovasc
373
Imaging. 2009;2(8):1024-1030. doi:10.1016/j.jcmg.2009.03.019. 4.
Restrepo M, Luffel M, Sebring J, et al. Surgical Planning of the Total Cavopulmonary
TE D
374 375
Connection: Robustness Analysis. Ann Biomed Eng. 2014;43(6):1321-1334.
376
doi:10.1007/s10439-014-1149-7. 5.
Tang E, Restrepo M, Haggerty C, et al. a Retrospective Cohort of 100 Fontan Connections:
EP
377
Relationship Between Geometric Features and Hemodynamics Outcomes. J Am Coll
379
Cardiol. 2013;61(10):E490. doi:10.1016/S0735-1097(13)60490-8.
380
6.
Cardiovasc Surg. 2013;148(4):1-10. doi:10.1016/j.jtcvs.2013.11.060.
382
384
Haggerty CM, Restrepo M, Tang E, et al. Fontan hemodynamics from 100 patient-specific cardiac magnetic resonance studies: A computational fluid dynamics analysis. J Thorac
381
383
AC C
378
7.
Restrepo M, Mirabella L, Tang E, et al. Fontan pathway growth: A quantitative evaluation of lateral tunnel and extracardiac cavopulmonary connections using serial cardiac
19
ACCEPTED MANUSCRIPT
385
magnetic resonance. Ann Thorac Surg. 2014;97(3):916-922.
386
doi:10.1016/j.athoracsur.2013.11.015. 8.
Kogon B. Is the extracardiac conduit the preferred Fontan approach for patients with
RI PT
387
univentricular hearts? The extracardiac conduit is the preferred Fontan approach for
389
patients with univentricular hearts. Circulation. 2012;126(21):2511-2515; discussion
390
2515. doi:10.1161/CIRCULATIONAHA.111.076398.
391
9.
SC
388
Khiabani RH, Restrepo M, Tang E, et al. Effect of flow pulsatility on modeling the hemodynamics in the total cavopulmonary connection. J Biomech. 2012;45(14):2376-
393
2381. doi:10.1016/j.jbiomech.2012.07.010.
394
10.
M AN U
392
Tang E, Restrepo M, Haggerty CM, et al. Geometric characterization of patient-specific total cavopulmonary connections and its relationship to hemodynamics. JACC Cardiovasc
396
Imaging. 2014;7(3):215-224. doi:10.1016/j.jcmg.2013.12.010.
397
11.
TE D
395
Duncan BW, Desai S. Pulmonary arteriovenous malformations after cavopulmonary anastomosis. Ann Thorac Surg. 2003;76(5):1759-1766. doi:10.1016/S0003-
399
4975(03)00450-8.
400
12.
Shah MJ, Rychik J, Fogel MA, Murphy JD, Jacobs ML. ORIGINAL ARTICLES : CARDIOVASCULAR Pulmonary AV Malformations After Superior Cavopulmonary
AC C
401
Connection : Resolution After Inclusion of Hepatic Veins in the Pulmonary Circulation.
402
1997;4975(96):0-3.
403 404
EP
398
13.
Kanter KR, Haggerty CM, Restrepo M, et al. Preliminary clinical experience with a
405
bifurcated Y-graft Fontan procedure - A feasibility study. J Thorac Cardiovasc Surg.
406
2012;144(2):383-389. doi:10.1016/j.jtcvs.2012.05.015.
20
ACCEPTED MANUSCRIPT
407
14.
Haggerty CM, Kanter KR, Restrepo M, et al. Simulating hemodynamics of the Fontan Ygraft based on patient-specific in vivo connections. J Thorac Cardiovasc Surg.
409
2013;145(3):663-670. doi:10.1016/j.jtcvs.2012.03.076.
410
15.
RI PT
408
Yang W, Feinstein J a, Shadden SC, Vignon-Clementel IE, Marsden AL. Optimization of a Ygraft design for improved hepatic flow distribution in the fontan circulation. J Biomech
412
Eng. 2013;135(1):011002. doi:10.1115/1.4023089.
413
16.
SC
411
Yang W, Chan FP, Reddy VM, Marsden AL, Feinstein J a. Flow simulations and validation for the first cohort of patients undergoing the Y-graft Fontan procedure. J Thorac
415
Cardiovasc Surg. 2015;149(1):247-255. doi:10.1016/j.jtcvs.2014.08.069.
416
17.
M AN U
414
Marsden AL, Bernstein AJ, Reddy VM, et al. Evaluation of a novel Y-shaped extracardiac Fontan baffle using computational fluid dynamics. J Thorac Cardiovasc Surg.
418
2009;137(2):394-403.e2. doi:10.1016/j.jtcvs.2008.06.043.
419
18.
TE D
417
Yang W, Feinstein J a., Marsden AL. Constrained optimization of an idealized Y-shaped baffle for the Fontan surgery at rest and exercise. Comput Methods Appl Mech Eng.
421
2010;199(33-36):2135-2149. doi:10.1016/j.cma.2010.03.012. 19.
difference for minimized total cavopulmonary power loss. J Biomech Eng.
423
2010;132(3):031012. doi:10.1115/1.4000954.
424 425
Dur O, DeGroff CG, Keller BB, Pekkan K. Optimization of inflow waveform phase-
AC C
422
EP
420
20.
Frakes DH, Smith MJT, Parks J, Sharma S, Fogel SM, Yoganathan AP. New techniques for
426
the reconstruction of complex vascular anatomies from MRI images. J Cardiovasc Magn
427
Reson. 2005;7(2):425-432. doi:10.1081/JCMR-200053637.
21
ACCEPTED MANUSCRIPT
428
21.
Frakes DH, Conrad CP, Healy TM, et al. Application of an adaptive control grid interpolation technique to morphological vascular reconstruction. IEEE Trans Biomed
430
Eng. 2003;50(2):197-206. doi:10.1109/TBME.2002.807651.
431
22.
Frakes D, Smith M, de Zélicourt D, Pekkan K, Yoganathan A. Three-Dimensional Velocity Field Reconstruction. J Biomech Eng. 2004;126(6):727. doi:10.1115/1.1824117.
432
23.
Kartik S. Sundareswaran, David H. Frakes, Mark A. Fogel, Dennis D. Soerensen, John N.
SC
433
RI PT
429
Oshinski APY. Optimum Fuzzy Filters for Phase Contrast MRI Segmentation. J Magn Reson
435
Imaging. 2009;29(1):155-165. doi:10.1002/jmri.21579.Optimum.
436
24.
M AN U
434
Zélicourt D De, Ge L, Wang C, Sotiropoulos F, Gilmanov A, Yoganathan A. Flow simulations in arbitrarily complex cardiovascular anatomies - An unstructured Cartesian
438
grid approach. Comput Fluids. 2009;38(9):1749-1762.
439
doi:10.1016/j.compfluid.2009.03.005.
440
25.
TE D
437
Schiavazzi DE, Kung EO, Marsden AL, et al. Hemodynamic effects of left pulmonary artery stenosis after superior cavopulmonary connection: A patient-specific multiscale modeling
442
study. J Thorac Cardiovasc Surg. 2015;149(3):689-696.e3.
443
doi:10.1016/j.jtcvs.2014.12.040. 26.
Fontan procedure: Chronic passive congestion, cardiac cirrhosis, hepatic adenoma, and
445
hepatocellular carcinoma. J Thorac Cardiovasc Surg. 2005;129(6):1348-1352.
446
doi:10.1016/j.jtcvs.2004.10.005.
447 448 449
Ghaferi A a., Hutchins GM. Progression of liver pathology in patients undergoing the
AC C
444
EP
441
27.
Asrani SK, Asrani NS, Freese DK, et al. Congenital heart disease and the liver. Hepatology. 2012;56(3):1160-1169. doi:10.1002/hep.25692.
22
ACCEPTED MANUSCRIPT
450
28.
Mertens L, Hagler DJ, Sauer U, Somerville J, Gewillig M. Protein-losing enteropathy after the Fontan operation: an international multicenter study. PLE study group. J Thorac
452
Cardiovasc Surg. 1998;115(5):1063-1073. doi:10.1016/S0022-5223(98)70406-4. 29.
Annu Rev Biomed Eng. 1999;1:299-329. doi:10.1146/annurev.bioeng.1.1.299.
454 455
Wootton DM, Ku DN. Fluid mechanics of vascular systems, diseases, and thrombosis.
30.
Rayz VL, Boussel L, Ge L, et al. Flow residence time and regions of intraluminal thrombus
SC
453
RI PT
451
deposition in intracranial aneurysms. Ann Biomed Eng. 2010;38(10):3058-3069.
457
doi:10.1007/s10439-010-0065-8.
M AN U
456
458 459
AC C
EP
TE D
460
23
ACCEPTED MANUSCRIPT
461
Figure Legends
462
Figure 1: (a) TCPC with an extracardiac baffle. (b) TCPC with a Y-Graft. (SVC: superior vena cava.
464
LPA: left pulmonary artery. RPA: right pulmonary artery.)
465
RI PT
463
Figure 2: Y-Graft hemodynamic performance. (a) Pulmonary and hepatic flow distribution (PFD
467
and HFD) deviations. (b) Effect of LPA stenosis on index power loss (iPL). (c) Distribution of TCPC
468
resistance. (d) Distribution of HFD deviation.
M AN U
SC
466
469
Figure 3: Representative Y-Graft cases with HFD noted. Streamlines colored by origin: IVC(blue),
471
SVC(green) and LSVC(red). Percentages shown are percent of hepatic flow to the left pulmonary
472
artery.
473
TE D
470
Figure 4: Hemodynamic Comparison of the Y-graft and LT/ECC. (a) Indexed power loss. (b) HFD
475
deviation from 50%. (c) Particle residence time. (d) Distribution of iPL. (e) Distribution of HFD
476
deviation. (f) Effect of LPA stenosis on iPL.
478 479
AC C
477
EP
474
Central Picture: Fontan Y-graft hepatic flow distribution; streamlines colored by inlet vessel.
24
ACCEPTED MANUSCRIPT
0.67 ± 0.12
0.66 ± 0.20
0.19
Age (years)
3.65 ± 1.14
3.68 ± 2.49
0.17
70%
PA Stenosis (%)*
33.99 ± 12.60
Systemic Venous Flow (L/min)* Pulmonary Flow Distribution (% LPA)
Pulsatility Index (SVC)
492
0.19
26.82 ± 10.69
0.02
1.57 ± 0.59
<0.01
46.01 ± 15.00
46.00 ± 16.78
0.99
13.13 ± 7.91
13.40 ± 10.60
0.78
0.66 ± 0.65
0.93 ± 0.72
0.14
0.42 ± 0.12
0.76
0.51 ± 0.50
EP
BSA: body surface area; PA: pulmonary artery; PFD: pulmonary flow distribution. Systemic venous return is defined as the sum of all TCPC inlet flow rates. Pulmonary flow distribution is the percent of total inlet flow that exits through the left pulmonary artery. An “*” represents statistically significant differences.
AC C
491
53%
2.07 ± 0.61
TE D
PFD Deviation from 50% (percentage points) Pulsatility Index (IVC)
490
p-value
BSA (m2)
Gender (male)
485 486 487 488 489
RI PT
Table 1: Averaged clinical data for all patients separated by graft type. LT/ECC Y-Graft (n=30) (n=30)
SC
481 482 483 484
Table 1:
M AN U
480
493 494
25
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT