- Email: [email protected]
Extracorporeal membrane oxygenator support in infants with systemic-pulmonary shunts
Accepted Manuscript Extra-corporeal membrane oxygenator support in infants with systemic-pulmonary shunts Phil Botha, PhD, FRCS, Shriprasad R. Deshpande, MBBS MS, Michael Wolf, MD, Michael Heard, RRT, Bahaaldin Alsoufi, MD, Brian Kogon, MD, Kirk Kanter, MD PII:
S0022-5223(16)30086-1
DOI:
10.1016/j.jtcvs.2016.03.075
Reference:
YMTC 10503
To appear in:
The Journal of Thoracic and Cardiovascular Surgery
Received Date: 15 July 2015 Revised Date:
26 March 2016
Accepted Date: 30 March 2016
Please cite this article as: Botha P, Deshpande SR, Wolf M, Heard M, Alsoufi B, Kogon B, Kanter K, Extra-corporeal membrane oxygenator support in infants with systemic-pulmonary shunts, The Journal of Thoracic and Cardiovascular Surgery (2016), doi: 10.1016/j.jtcvs.2016.03.075. 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
Title Page Title: Extra-corporeal membrane oxygenator support in infants with systemic-pulmonary shunts. Running Head: ECMO in Infants with SP Shunts
RI PT
Authors: Phil Botha PhD, FRCS, Shriprasad R. Deshpande MBBS MS, Michael Wolf MD, Michael Heard RRT, Bahaaldin Alsoufi MD, Brian Kogon MD and Kirk Kanter MD . Institutions and Affiliations: Children’s Healthcare of Atlanta, Emory University, Atlanta, Georgia.
Corresponding Author:
EP
TE D
Shriprasad R. Deshpande Assistant Professor Department of Pediatrics Division of Pediatric Cardiology Emory University School of Medicine Children's Healthcare of Atlanta 1405 Clifton Road Atlanta, GA, USA 30322-1101 Email: [email protected] Phone: 404-694-7739 Fax : 770-488-9480
M AN U
Word Count: 3441 (manuscript word count)
SC
Keywords: Extracorporeal membrane oxygenation, ECMO; Shunts (Aorto-pulmonary); Congenital heart disease; CHD, Univentricular heart; Pediatric.
Disclosure : None of the authors have any conflicts of interest to disclose. Funding : No funding was provided or utilized for the conduct of this study
AC C
Abbreviations : ECMO : Extracorporeal membrane oxygenation SP shunt : systemic-pulmonary shunt BT shunt : modified Blalock- Taussig shunt eCPR : Extra-corporeal cardiopulmonary resuscitation CVVH : Continuous veno-venous hemofiltration
ACCEPTED MANUSCRIPT
Abstract Background
RI PT
Management of a patent systemic-pulmonary shunt and the resultant run-off whilst on extracorporeal membrane oxygenation (ECMO) varies amongst institutions. We have employed a strategy of increased flow without surgical reduction of the shunt diameter, and describe our results with this strategy. Methods
SC
Database review of 169 successive veno-arterial ECMO runs in infants and neonates. ECMO flow, time taken to achieve lactate clearance, normal pH and negative fluid balance was
M AN U
contrasted in patients with and without shunts. Results
Between 2002 and 2013, 51 of 169 (30.2%) infants had a shunt in situ when ECMO was initiated. Significantly higher ECMO flows were maintained in the shunt group (161±43 ml/kg/min vs. 134±41 ml/kg/min, P<0.001). Infants with shunts had significantly higher preECMO and peak lactates (12.4±5.6 mmol/l vs. 10.0±6.3 mmol/l, p<0.05, and 13.7±4.9 mmol/l
TE D
vs. 11.6±5.5 mmol/l, p<0.02 respectively) and clearance required a longer period of support (median 28.8, 16.1 – 63.3h vs. 17.5, 10.8 – 34.5h, p<0.001). Although the absolute rate of lactate clearance was not significantly different between groups (Median 0.46, 0.12 – 0.72 mmol/l/h vs. 0.48, 0.22 – 0.86mmol/l/h, p = 0.139) the presence of a shunt, neonatal age, peak lactate, extra-
EP
corporeal cardiopulmonary resuscitation and the use of hemofiltration on ECMO significantly predicted the rate of clearance. Survival to hospital discharge was similar in the shunt and non-
AC C
shunt groups (49.0% vs. 48.3%, p=0.932). Conclusions
A strategy of increased ECMO flow without surgically restricting shunt diameter appears successful in providing circulatory support in the majority of patients with systemic-pulmonary shunts. Equivalent survival suggests that routine surgical reduction of shunt diameter is not indicated.
Word Count: 249
ACCEPTED MANUSCRIPT
Introduction
2
The circulation in the presence of a surgical systemic-pulmonary (SP) shunt is one of inherent
3
instability, particularly in the setting of single ventricle physiology. Fluctuations in either the
4
systemic or pulmonary vascular tone can result in a profound and rapid alteration of systemic
5
perfusion. In the presence of systemic diastolic run-off into the low resistance pulmonary bed,
6
any reduction in the systemic perfusion pressure can lead to a rapid decompensation of
7
myocardial perfusion and impaired cardiac output in return. In addition to a tendency to periods
8
of pulmonary over-circulation, acute thrombosis of the shunt will similarly precipitate a
9
profound collapse of the cardiovascular system. For these reasons, the clinician caring for
SC
RI PT
1
patients with shunts is not infrequently faced with a need to support the collapsed circulation
11
using extra-corporeal membrane oxygenation (ECMO).
12
Initial reports of ECMO for post-cardiotomy circulatory support considered patients with
13
systemic-pulmonary shunts unsuitable. The resultant run-off, increased pulmonary flow and
14
therefore pulmonary venous return were thought to preclude successful support using this
15
modality[1,2]. Complete occlusion of the shunt by means of a surgical clip at the time of ECMO
16
was shown to lead to unacceptable mortality rates and many centers have therefore adopted a
17
strategy of increased flow to compensate for shunt run-off[3]. It has been suggested that partial
18
occlusion of the shunt during ECMO may reduce run-off into the pulmonary bed, resulting in
19
improved perfusion pressure and hasten resolution of the malperfusion state[4]. Our institution
20
has pursued a strategy of liberal flow on ECMO without applying any limitation to shunt run-off.
21
Here, we review our results using this strategy in terms of resolution of the malperfusion state
22
and other outcomes.
EP
TE D
M AN U
10
AC C
23 24
Methods
25
Our prospectively collected ECMO database was interrogated to identify all patients that
26
underwent ECMO for cardiac indications between January 2002 and October 2013. This data
27
was supplemented by chart review and examination of the electronic record of laboratory
28
investigations. Neonates and infants that underwent veno-arterial ECMO whilst having an
29
systemic-pulmonary shunt in situ (Shunt group) were contrasted to all infants and neonates
30
without shunts during the same period (Non-shunt group). Patients that underwent the first stage
31
Norwood procedure with a Sano RV-PA conduit were not included in the shunt group. In
ACCEPTED MANUSCRIPT
32
patients that underwent pre- and post-cardiotomy ECMO, only the post-operative episode was
33
analyzed.
34 Site of ECMO cannulation was dictated by the clinical situation. Cannulation was performed via
36
central cannulation (reopening of the sternotomy incision ) in the early post-operative period and
37
via a neck incision (jugular and carotid vessels ) in those requiring ECMO after the two-week
38
post-operative period.
39
With cannulation via the right carotid artery, we aim to place the tip of the arterial cannula at the
40
junction of the innominate artery and the aortic arch. This does produce the possibility of the
41
systemic end of the shunt being occluded by the cannula. We do however routinely continue
42
ventilation in this setting, and a loss of end-tidal CO2 would demonstrate that the shunt is being
43
occluded by the cannula. We have not observed this phenomenon. We further confirm the
44
patency with routine echocardiogram during the ECMO as well as during the weaning trials.
45
Again, if there is any concern about the cannula placement /shunt occlusion, the cannulae are
46
repositioned.
47
Similarly for patients with an open chest post sternotomy, direct central cannulation was
48
performed which involved cannulation of the right atrium (venous) and the aorta (arterial).
49
ECMO was delivered using a standard circuit (S-97-E Tygon tubing, Medtronic, Minneapolis,
50
MN), oxygenator (pediatric Quadrox D ©Maquet, Rastatt, Germany) and roller pump (Century,
51
Mesa, AZ) during this period. ECMO flows were increased to around 100ml/kg/min in non-
52
shunted patients and up to 200ml/kg/min in patients with systemic-pulmonary shunts. Flow rate
53
was adjusted according to pre-membrane oxygen saturation and blood pressure thereafter.
SC
M AN U
TE D
EP
AC C
54
RI PT
35
55
Serum lactic acid was measured using a regularly calibrated point-of care analyzer (iSTAT,
56
Abbott Point of care) and is presented as mmol/l. The level of serum lactate prior to
57
commencement of ECMO and the peak serum lactate measured during ECMO support was
58
documented. Typically, serum lactates were measured approximately two hourly in the early
59
phase after institution of ECMO and the frequency of sampling adjusted according to clinical
60
need thereafter. Time to lactate clearance was defined as the time elapsed between the institution
61
of ECMO and the serum lactate reaching a value below 2.0mmol/l (the upper limit of normal in
62
our laboratory). Rate of lactate clearance was calculated as peak lactate – 2.0, divided by the
ACCEPTED MANUSCRIPT
time elapsed between initiating ECMO and the first lactate level below 2.0mmol/l recorded. For
64
analysis of variables affecting the rate of lactate clearance, the range of clearance rates was
65
divided into equal quartiles and entered into univariate and multivariate logistic regression
66
models. For multivariate analysis, all variables with p values less than 0.10 in univariate analysis
67
were entered into the multinomial logistic regression model in a stepwise fashion. This analysis
68
considered the highest quartile of lactate clearance as the reference category and parameter
69
estimates for the contrast with the lowest quartile are reported.
70
Arterial blood gas analysis was performed at hourly intervals initially after institution of ECMO
71
using the same point-of-care analyzer as for lactic acid measurements. The fluid balance was
72
recorded in the electronic record throughout and analyzed in 12-hour intervals from 07:00 to
73
19:00 and 19:00 to 07:00. The time at conclusion of the first 12-hour interval after the initiation
74
of ECMO during which a negative fluid balance was achieved for that 12-hour period was taken
75
as the time at which a negative fluid balance was attained for the analysis. Total fluid balance
76
was calculated as total measured intake minus total measured fluid loss, not accounting for
77
insensible losses.
M AN U
SC
RI PT
63
78
Statistical analyses we're performed using the SPSS v.21. Parametric continuous variables are
80
presented as mean ± standard deviation and analyzed using the student’s t-test. Non-parametric
81
variables are presented as median followed by the 25th - 75th percentiles and analyzed using the
82
Wilcoxon rank-sum test or Mann-Whitney test where indicated. In all analyses, a p-value less
83
than 0.05 was considered statistically significant. Our institutional review board has reviewed the
84
study protocol and has waived the need for individual informed consent. All authors have had
85
full access to the data and have read and agreed to this final version of the manuscript.
87
EP
AC C
86
TE D
79
88
Results
89
During the period January 2002 to October 2013, 218 patients underwent 224 ECMO runs for
90
cardiac indications. Of these, 169 patients were neonates or infants at the time of commencing
91
ECMO ; 51 with an systemic-pulmonary shunt and the remaining 118 representing the non-shunt
92
group. One patient with a diagnosis of Pulmonary atresia/ ventricular septal defect with a
93
modified Blalock - Taussig (BT) shunt placed at a prior admission underwent veno-venous
ACCEPTED MANUSCRIPT
ECMO for a suspected occluded shunt and was therefore not included in the analysis.
95
Demographic variables in the shunt and non-shunt groups are contrasted in Table 1. Age was
96
not significantly different in the shunt and non-shunt groups (median 15 and 19 days
97
respectively, p=0.384), but infants with shunts weighed significantly less (3.3 vs. 3.6kg,
98
p<0.010). A greater proportion of shunt patients underwent ECMO during the immediate post-
99
operative admission (94.1% vs. 76.3%, p<0.005).
100
RI PT
94
ECMO conduct
102
ECMO was instituted in the operating room after failure to wean from cardiopulmonary bypass
103
or for marginal hemodynamics after weaning in 20 patients in the non-shunt group (16.9%) and 6
104
patients in the shunt group (11.9%, p=0.489). The remainder were cannulated in the cardiac
105
intensive care unit. Of the ones cannulated in the CICU, the common indication was cardiac
106
arrest. These patients were therefore resuscitated using the extra-corporeal cardiopulmonary
107
resuscitation (ECPR) pathway. A greater proportion of shunt patients underwent (ECPR)
108
protocol (72.5% vs. 47.5%, p<0.010). As shown in table 1, other indications included
109
cardiovascular instability, hypoxia, respiratory arrest and concern for shunt thrombosis. This
110
distribution was comparable in the shunt and non-shunt groups. Decision regarding
111
cardiovascular instability was usually based on increasing needs for inotropic and vasopressor
112
support, increase in lactic acidosis, as well as other measures of low cardiac output such as low
113
NIRS, poor perfusion, narrow pulse pressure with tachycardia. The decision was made by the
114
intensivist in conjunction with the primary surgeon. Patients who were cannulated for the
115
concern of shunt thrombosis (5/51) underwent a diagnostic cardiac catheterization /angiography
116
to visualize the shunt patency after stabilization. Shunt patency was established in each one of
117
these patients via angiography and or intervention as needed within 24 hours of initiation of
118
ECMO. All five of these patients therefore had a patent shunt through the rest of their ECMO
119
course.
M AN U
TE D
EP
AC C
120
SC
101
121
Cannulation was performed via the neck vessels in 10 patients in the shunt group (19.6%) and 33
122
patients in the non-shunt group (28.0%, p=0.336), and via the open chest in the remainder.
123
Significantly higher ECMO flows were maintained in the shunt group at 4 hours (median 161±43
ACCEPTED MANUSCRIPT
124
ml/kg/min vs. 134±41 ml/kg/min, p<0.001) and at 24 hours (median 148±43 ml/kg/min vs.
125
124±38 ml/kg/min, P<0.01).
126 Lactate Clearance
128
Serum lactic acid levels were significantly higher in patients with SP shunts immediately prior to
129
going onto ECMO than in patients without shunts (12.4±5.6 mmol/l vs. 10.0±6.3 mmol/l,
130
p<0.05). Similarly lactate levels peaked significantly higher in patients with shunts (13.7±4.9
131
mmol/l vs. 11.6±5.5 mmol/l, p<0.02). Five patients died or were de-cannulated before lactate
132
clearance could be achieved (3 in the shunt group and 2 in the non-shunt group). A significantly
133
greater proportion of patients in the shunt group required more than 48 hour to achieve lactate
134
clearance (31.4% vs. 12.7%, p=0.004). Median time to achieve lactate clearance was
135
significantly longer in patients with systemic-pulmonary shunts 28.8, 16.1 – 63.3h vs. 17.5, 10.8 –
136
34.5h, p<0.001 . Figure 1 demonstrates graphically the pre-ECMO and peak lactic acid levels, and
137
median time to lactate clearance in patients with and without shunts requiring ECMO. Although
138
the rate of lactate clearance appeared to be lower in shunt patients, the median difference did not
139
reach statistical significance (Median 0.46, 0.12 – 0.72 mmol/l/h vs. 0.48, 0.22 – 0.86mmol/l/h, p
140
= 0.139). Results of univariate and multivariate analysis for factors affecting lactate clearance are
141
listed in Table 2. In the multivariate model, peak serum lactic acid during ECMO, ECPR (as
142
opposed to elective institution of ECMO), neonatal age and the use of continuous veno-venous
143
hemofiltration (CVVH) were significant predictors of lactate clearance. The presence of a
144
systemic-pulmonary shunt was significantly predictive of belonging to the lowest quartile for
145
rate of lactate clearance but the interaction term size of shunt/body weight was not a statistically
146
significant predictor. There was no relationship between the ECMO flow rate and clearance of
147
lactate and this was independent of the presence or absence of SP shunt.
SC
M AN U
TE D
EP
AC C
148
RI PT
127
149
Acidemia/ Fluid balance
150
The median lowest arterial blood pH recorded in the period immediately prior to ECMO
151
cannulation or whilst on ECMO was 7.120, 7.022 – 7.225 in the shunt group and 7.183, 7.054 –
152
7.244 in the non-shunt group. This difference did not reach statistical significance on the Mann-
153
Whitney U-test (p=0.074). Time taken to achieve an arterial blood pH of greater than 7.350 on
154
ECMO also did not differ significantly between the shunt and non-shunt groups (median 0.62,
ACCEPTED MANUSCRIPT
0.33 – 1.07h vs. 0.68, 0.23 – 1.15h, p=0.822.) Data on fluid balance during ECMO was available
156
for 24 patients in the shunt group and 70 patients in the non-shunt group. Median time to achieve
157
a 12-hour period of overall negative fluid balance was 47.0, 34.6 – 59.2h vs. 31.9, 19.1 – 49.1h
158
in the shunt and non-shunt groups respectively (p=0.069). The overall fluid balance was not
159
significantly different in shunt and non-shunt groups, irrespective of normalization for the
160
duration of ECMO (+24.6±657.3 ml/day vs. +48.9±888.4 ml/day, p=0.385).
RI PT
155
161 Norwood Stage 1
163
A pre-specified subgroup analysis was performed using the cohort of patients that required
164
ECMO support in the post-operative period after stage 1 Norwood procedure. This comprised a
165
group of 44 patients, 16 (36%) having received modified BT shunts and 28 (64%) having
166
received Sano (right ventricle to pulmonary artery) conduits. The size of the BT shunt was 3mm
167
in 3 patients and 3.5mm in 14 patients. Table 3 summarizes demographic and ECMO variables
168
contrasted in these two groups. Patients in these two groups had similar mean weights and a
169
similar proportion underwent ECMO cannulation as part of an ECPR protocol. Pre-ECMO
170
lactates and peak lactates were not significantly different between the two groups. Norwood
171
stage 1 patients with BT shunts had a lower rate of lactate clearance on ECMO than patients with
172
Sano conduits (0.29, 0.08 – 0.56 vs. 0.47, 0.28 – 0.82 mmol/l/hr, Mann-Whitney p=0.052).
173
TE D
M AN U
SC
162
Outcomes
175
Overall median duration of ECMO for infants and neonates was 119.3 hours (Range 3.3 - 651.7
176
hours). Duration of ECMO support was significantly greater in the non-shunt group (131.0, 78.4
177
- 195h vs. 77.6, 41.7 - 175h, p=0.021). Surrogate measures of low cardiac output or malperfusion
178
state such as renal insufficiency were also assessed as secondary endpoints. In addition to the
179
fluid balance mentioned above (which reflected the urine output), serum creatinine as well as
180
need for CVVH post-ECMO was compared. serum creatinine values at decannulation or at the
181
time of discharge. Only two patients needed CVVH post-ECMO , both were on CVVH during
182
the ECMO run. Both these patients belonged to the shunt group. Due to a very small number
183
(2/169 patients), further analysis related to persistent renal function post- ECMO , was not
184
performed.
AC C
EP
174
ACCEPTED MANUSCRIPT
For the entire cohort of patients receiving ECMO support over this time period, survival to
186
ECMO de-cannulation was 71.1% and survival to hospital discharge was 50.9%. Survival for
187
infants and neonates overall was 48.5%, and not significantly different in shunt and non-shunt
188
groups (49.0% vs. 48.3%, p=0.932). In the cohort of patients post stage 1 Norwood, survival was
189
nearly identical, being 71.1% survival to ECMO de-cannulation and 46.7% survival to discharge.
RI PT
185
190 Discussion
193
Palliation of single ventricle patients with a staged approach is now the standard accepted
194
strategy [5]. In conditions with inadequate pulmonary blood flow and those requiring single
195
ventricle palliation, the systemic-pulmonary shunt still represents in many patients the first
196
surgical step (or part thereof). The period from the creation of the shunt until definitive repair, or
197
the substitution of the systemic-pulmonary shunt with a cavo-pulmonary shunt, remains the most
198
critical phase in terms of risk for mortality and morbidity [6]. The shunt-dependent circulation is
199
one characterized by lability, as the synthetic conduit used for the majority of such shunts does
200
not allow for the usual degree of cardiovascular auto-regulation. Creating a connection between
201
the usually serial pulmonary and systemic circulations can lead to rapid and profound swings in
202
perfusion in favor of either vascular bed. Patient populations undergoing SP shunts have changed
203
considerably over the past 70 years[7,8] but the rate of in-hospital mortality has changed little
204
from 8.1% in the earliest reports[9] to rates of around 4-10% in the higher risk contemporary
205
populations [8,10,11]. Major morbidity also remains common after SP shunts, with re-
206
intervention required in 7.3% and around 3-6% requiring ECMO support[11,12]. We have
207
analyzed a consecutive series of 169 infants and neonates requiring mechanical circulatory
208
support in the form of veno-arterial ECMO, to ascertain the impact of the presence of an
209
systemic- pulmonary shunt during ECMO support on their outcomes.
M AN U
TE D
EP
AC C
210
SC
191 192
211
In early series of pediatric post-cardiotomy ECMO, patients with shunts were considered not to
212
be suitable candidates because of the limitations shunt run-off would impose on systemic
213
perfusion [1,2]. Jaggers and colleagues reported outcomes in 9 patients with shunts requiring
214
ECMO [3]. The authors occluded the shunt surgically at the time of going on to ECMO in 4
215
patients, but had no survivors in this group. In contrast, the mortality rate in 5 patients where the
ACCEPTED MANUSCRIPT
shunt was left patent and the ECMO flow increased to compensate for shunt run-off was
217
considerably lower at 20%. More recently, Allan and colleagues reported 44 patients that
218
underwent ECMO following single ventricle palliation with a systemic to pulmonary arterial
219
shunt [4]. The authors maintained shunt patency wherever possible, but in patients demonstrating
220
decreased systemic perfusion with delayed clearance of lactate despite increasing ECMO flow, 1
221
or more surgical clips were placed to restrict the shunt diameter and reduce run-off. This was
222
required in 17 of 44 patients (39%). Overall survival to hospital discharge was 48%, with 20% of
223
survivors requiring shunt clipping, as opposed to 61% of non-survivors. In the present series we
224
have found shunt patients requiring ECMO to have lower body weights, to have required ECMO
225
predominantly in the post-operative period and more commonly as part of extra-corporeal
226
resuscitation for cardiovascular collapse. Significantly higher ECMO flows were maintained in
227
the shunt group and no surgical reduction of shunt diameter was undertaken. Survival to
228
discharge in all infants and neonates requiring ECMO in our series was comparable to these
229
previous publications and not affected by the presence of a systemic-pulmonary shunt.
M AN U
SC
RI PT
216
230
As a surrogate measure of tissue oxygen delivery and anaerobic metabolism, the serum lactic
232
acid at various post-operative time-points has been shown to predict outcomes in pediatric
233
cardiac surgery [13-15]. Several factors will determine the level of serum lactate in an infant
234
requiring ECMO. The duration and severity of the low cardiac output/ hypoxemic state
235
prompting the need for ECMO will be major determinants, but other factors including duration
236
of cardiopulmonary bypass, circulatory arrest and/ or regional perfusion have also been shown to
237
correlate with early post-operative serum lactate [16,17]. As patients in the shunt group in our
238
series were usually in the peri-operative period and more likely to have required ECMO after
239
cardiac arrest and therefore as part of an eCPR protocol, we found not surprisingly that pre-
240
ECMO serum lactic acid was significantly higher in this group. The determinants of the rate of
241
lactate clearance in a patient on ECMO are also multiple. Whether the precipitant of
242
cardiovascular collapse has resolved will have a profound effect, as will the presence of residual
243
uncorrected anatomical abnormalities. The adequacy of organ perfusion once on ECMO will
244
determine the level of ongoing lactate production and the rate at which the existing pool will be
245
metabolized. Although the rate of lactate clearance in shunted patients in our series appeared to
246
be lower, the median difference in the rate of clearance on ECMO did not reach statistical
AC C
EP
TE D
231
ACCEPTED MANUSCRIPT
significance. On multivariate analysis peak serum lactic acid during ECMO, ECPR, neonatal age
248
and the use of continuous veno-venous hemofiltration (CVVH) were significant predictors of the
249
rate of lactate clearance. The presence of a systemic-pulmonary shunt was also a significant
250
predictor of belonging to the lowest quartile for rate of lactate clearance but the relative size of
251
the shunt did not appear significant. It has been a concern that achieving and maintaining higher
252
ECMO flows in patients with a patent systemic-pulmonary shunt may be at the expense of a
253
greater need for volume administration, and therefore delay the resolution of fluid sequestration.
254
Median time in hours to achieve a 12-hour period of overall negative fluid balance was 47.0,
255
34.6 – 59.2h vs. 31.9, 19.1 – 49.1h in the shunt and non-shunt groups respectively. Although
256
many factors could have influenced this trend, including lower renal perfusion and a higher
257
requirement for CVVH in the shunt group, the difference did not reach statistical significance
258
(p=0.069). The lower number of patients in which all the required datapoints were available,
259
cautions the possibility of a type-2 error in this analysis.
M AN U
SC
RI PT
247
260
We also analyzed as a more homogenous subgroup, all patients that required ECMO after stage 1
262
Norwood contrasting those with systemic-pulmonary shunts with those with RV-PA (Sano)
263
shunts. Although pre-ECMO and peak lactates were not significantly different between groups,
264
patients with BT shunts had a significantly lower rate of lactate clearance than patients with Sano
265
conduits. Polimenakos and colleagues reported on 20 functional single ventricle patients
266
supported with ECMO post cardiotomy[18]. They left the shunt patent in all cases in the hope of
267
minimizing pulmonary vascular endothelial injury and utilized higher pump flows
268
(>150ml/kg/min) to compensate for shunt runoff. Vaso-active-inotropic support was continued
269
and 20ppm Inhaled Nitric oxide administered in all cases. Twelve of these patients had
270
pulmonary blood flow established via a Sano shunt, which would likely have significantly
271
smaller impact on flow requirement. Survival to discharge was 57%. Ravishankar and colleagues
272
also document a survival to discharge of 39 % in 36 neonates rescued by ECMO following stage
273
1 palliation[19]. They note that the shunt was left patent (modified BT shunt in the majority of
274
cases), requiring flows between 150-200ml/kg/min to compensate for shunt run-off. Similarly,
275
the analysis from the ELSO registry [20] showed that for patients HLHS after stage I palliation,
276
needing ECMO, 59% survived the decannulation while 31% survived to discharge. The risk
277
factors associated with death were Black race, lower weight and longer duration of ventilator
AC C
EP
TE D
261
ACCEPTED MANUSCRIPT
support prior to ECMO. However, they were unable to comment on the type of shunt (SP shunt
279
versus a Sano), other surgical details (such as clipped shunts) or patient acuity prior to ECMO. In
280
our experience, despite a slower clearance of lactate in patients with systemic-pulmonary shunts,
281
survival to discharge was not affected irrespective of the shunt type and was uniformly better
282
(48%) than the ELSO registry report.
RI PT
278
283 284 Limitations
286
The study is inherently limited in generalizability by its retrospective nature and sample size.
287
Although a standardized protocol for the management of all patients on ECMO was followed,
288
this will have undergone some evolution over the 11year time-course of the study.
M AN U
SC
285
289 Conclusions
291
Mortality remains common in Infants requiring support by Extra-corporeal membrane
292
oxygenation. Accepting the limitations of mortality as an endpoint, the presence of a systemic-
293
pulmonary shunt during ECMO support does not appear to have an adverse effect on outcome, if
294
ECMO flow is appropriately increased to compensate for shunt run-off. Although patients with a
295
patent systemic-pulmonary shunt appear to have a delay in the resolution of the malperfusion
296
state on multivariate analysis, routine surgical reduction of shunt diameter to reduce shunt run-
297
off cannot be supported based on our data.
AC C
EP
TE D
290
ACCEPTED MANUSCRIPT
Figure Legend Non-shunt group (N=118)
p
Age (days)
15 (8-38)
19 (7-109)
N.S.
Weight (kg)
3.3 (2.7-3.8)
3.6 (2.9 - 5.4)
0.010
Gender: Male
69 (58.5%)
27 (52.9%)
N.S.
Diagnosis: HLHS
10 (19.6%)
23 (19.5%)
PA/IVS
11 (21.6%)
4 (3.4%)
Other Single ventricle
22 (43.1%)
TAPVR
0
DCM/ Myocarditis
0
Truncus arteriosus Other Post-operative
ECPR Shunt diameter: 3.0mm 3.5mm
SC 11(9.3%) 9 (7.6%)
0
8 (6.8%)
7 (13.7%)
37 (31.4%)
48 (94.1%)
90 (76.3%)
0.005
16 (31.4%)
28 (23.7%)
N.S.
37 (72.5%)
56 (47.5%)
0.003
8 (15.7%) 41 (80.4%) 2 (3.9%)
EP
4.0mm
15 (12.7%)
1 (2.0%)
TE D
Norwood stage 1
11 (9.3%)
M AN U
TGA
RI PT
Shunt Group (N=51)
Indication for ECMO: Unable to wean CPB
13 (11.01%)
Cardiac Arrest
19 (37.3%)
37 (31.4%)
CV Instability
16 (31.4%)
38+7 (38.1%)
Hypoxia
6 (11.8%)
11 (9.3%)
Respiratory arrest
7 (6.1%)
3 (2.5%)
AC C
2 (3.9%)
Concern for shunt 5 (9.8%) Thrombosis
0
Table 1. Patient demographics. The shunt group comprises neonatal and infant patients requiring VAECMO support during the study period. N.S. - Not significant, HLHS- Hypoplastic left heart syndrome, PA/IVS - Pulmonary atresia with intact ventricular septum, TAPVR - Total anomalous pulmonary venous return, DCM- Dilated cardiomyopathy, TGA, Transposition of the great arteries, ECPR - Extracorporeal cardiopulmonary resuscitation, CPB - Cardiopulmonary bypass, CV – cardiovascular.
Multivariate Coefficient
95% CI
Neonatal age
0.013
0.029
4.291
1.160-15.866
Weight (kg)
0.111
ECPR
<0.001
SP Shunt
0.174
0.019
4.926
Shunt diameter / weight (mm/kg)
0.409
Pre-ECMO lactate
0.012
Peak lactate
<0.001
<0.001
CVVH
0.006
0.050
Stage 1 Norwood
0.344
CPB time (min)
0.148
Crossclamp (min)
0.968
Major bleeding
0.158
Neurological event
0.595
SC
RI PT
Multivariate p
TE D
Univariate p
M AN U
ACCEPTED MANUSCRIPT
1.303-18.681
0.729
0.634-0.839
0.280
0.079-0.999
AC C
EP
Table 2. Results of univariate and multivariate analyses for factors predictive of delayed lactate clearance. ECPR – Extracorporeal cardiopulmonary resuscitation, SP – Systemic-pulmonary, CVVH – continuous veno-venous hemofiltration, CPB- cardiopulmonary bypass.
ACCEPTED MANUSCRIPT
Sano group (N=28)
p
Weight (kg)
3.28 ±0.67
3.11 ±0.67
0.493
ECPR
9 (56%)
17 (61%)
0.954
Indication for ECMO: Unable to wean CPB
2 (13%)
2 (7%)
Cardiac Arrest
2 (13%)
12 (43%)
CV instability
5 (31%)
11 (39%)
Hypoxia
1 (6%)
2 (7%)
Respiratory arrest
0
Metabolic
2 (13%)
Pre ECMO Lactate (mmol/l) Peak Lactate (mmol/l) Rate of Lactate clearance (mmol/l/hr) Time to pH > 7.35 (hr)
Total fluid balance (ml/day)
SC 0
4 (25%)
0
13.0 (5.2)
11.4 (5.1)
0.362
13.2 (4.4)
13.1 (4.9)
0.407
0.29 (0.32)
0.47 (0.41)
0.052
0.63 (0.46)
0.71 (2.36)
0.464
45.7 (20.5)
28.6 (20.95)
0.185
2.0 (310.0)
0.129
TE D
Time to negative fluid balance (hours)
1 (4%)
M AN U
Shunt Thrombosis
RI PT
BT Shunt group (N=16)
-116.1 (430.9)
AC C
EP
Table 3. Comparison of ECMO conduct and metabolic normalization in first stage Norwoood patients undergoing ECMO, contrasted in the subset with systemic-pulmonary shunts and those with right ventricle to pulmonary artery conduits.
ACCEPTED MANUSCRIPT
20 18
*
*
Shunt No Shunt
16
RI PT
* - p <0.05
12
Rate of lactate clearance: (Median 0.46, 0.12 – 0.72 mmol/l/h vs. 0.48, 0.22 – 0.86 mmol/l/h, p = 0.139)
10
SC
Lactate (mmol/l)
14
8
4
* 2 0 0
5
10
15
20
M AN U
6
25
30
35
40
45
50
55
60
65
TE D
Time (h)
AC C
EP
Figure 1. Line plot depicting mean lactic acid levels immediately prior to ECMO support and peak lactate level on ECMO, as well as median time to achieve normal lactic acid level in systemic-pulmonary shunt and non-shunt groups. Error bars depict standard deviation for parametric variables and 25-75% interquartile range for time to normalization of lactate.
ACCEPTED MANUSCRIPT
References [1]
Kanter KR, Pennington G, Weber TR, Zambie MA, Braun P, Martychenko V.
RI PT
Extracorporeal membrane oxygenation for postoperative cardiac support in children. J Thorac Cardiovasc Surg 1987;93:27–35. [2]
del Nido PJ, Dalton HJ, Thompson AE, Siewers RD. Extracorporeal membrane
oxygenator rescue in children during cardiac arrest after cardiac surgery. Circulation
[3]
SC
1992;86:II300–4.
Jaggers JJ, Forbess JM, Shah AS, Meliones JN, Kirshbom PM, Miller CE, et al. Extracorporeal membrane oxygenation for infant postcardiotomy support: significance
[4]
M AN U
of shunt management. Ann Thorac Surg 2000;69:1476–83.
Allan CK, Thiagarajan RR, del Nido PJ, Roth SJ, Almodovar MC, Laussen PC. Indication for initiation of mechanical circulatory support impacts survival of infants with shunted single-ventricle circulation supported with extracorporeal membrane oxygenation. J Thorac Cardiovasc Surg 2007;133:660–7.
[5]
Williams JA, Bansal AK, Kim BJ, Nwakanma LU, Patel ND, Seth AK, et al. Two
TE D
thousand Blalock-Taussig shunts: a six-decade experience. Ann Thorac Surg 2007;84:2070–5–discussion2070–5. [6]
Fenton KN, Siewers RD, Rebovich B, Pigula FA. Interim mortality in infants with
7.
EP
systemic-to-pulmonary artery shunts. Ann Thorac Surg 2003;76:152–6–discussion156–
Blalock A. The Surgical Treatment of Malformations of the Heart. Jama 1945;128:189.
[8]
Heidari-Bateni G, Norouzi S, Hall M, Brar A, Eghtesady P. Defining the best practice
AC C
[7]
patterns for the neonatal systemic-to-pulmonary artery shunt procedure. J Thorac
Cardiovasc Surg 2014;147:869–873.e3.
[9]
de Leval MR, McKay R, Jones M, Stark J, Macartney FJ. Modified Blalock-Taussig shunt. Use of subclavian artery orifice as flow regulator in prosthetic systemic-
pulmonary artery shunts. J Thorac Cardiovasc Surg 1981;81:112–9. [10]
Myers JW, Ghanayem NS, Cao Y, Simpson P, Trapp K, Mitchell ME, et al. Outcomes of systemic to pulmonary artery shunts in patients weighing less than 3 kg: Analysis of shunt type, size, and surgical approach. J Thorac Cardiovasc Surg 2014;147:672–7.
ACCEPTED MANUSCRIPT
[11]
Petrucci O, O'Brien SM, Jacobs ML, Jacobs JP, Manning PB, Eghtesady P. Risk Factors for Mortality and Morbidity After the Neonatal Blalock-Taussig Shunt Procedure. Ann Thorac Surg 2011;92:642–52.
[12]
Guzzetta NA, Foster GS, Mruthinti N, Kilgore PD, Miller BE, Kanter KR. In-Hospital
RI PT
Shunt Occlusion in Infants Undergoing a Modified Blalock-Taussig Shunt. Ann Thorac Surg 2013;96:176–82. [13]
Siegel LB, Dalton HJ, Hertzog JH, Hopkins RA, Hannan RL, Hauser GJ. Initial
Intensive Care Med 1996;22:1418–23. [14]
SC
postoperative serum lactate levels predict survival in children after open heart surgery.
Charpie JR, Dekeon MK, Goldberg CS, Mosca RS, Bove EL, Kulik TJ. Serial blood
M AN U
lactate measurements predict early outcome after neonatal repair or palliation for complex congenital heart disease. J Thorac Cardiovasc Surg 2000;120:73–80. [15]
Schumacher KR, Reichel RA, Vlasic JR, Yu S, Donohue J, Gajarski RJ, et al. Rate of increase in serum lactate level risk-stratifies infants after surgery for congenital heart disease. J Thorac Cardiovasc Surg 2013.
[16]
Munoz R, Laussen PC, Palacio G, Zienko L, Piercey G, Wessel DL. Changes in whole
TE D
blood lactate levels during cardiopulmonary bypass for surgery for congenital cardiac disease: an early indicator of morbidity and mortality. J Thorac Cardiovasc Surg 2000;119:155–62. [17]
Algra SO, Schouten ANJ, van Oeveren W, van der Tweel I, Schoof PH, Jansen NJG, et
EP
al. Low-flow antegrade cerebral perfusion attenuates early renal and intestinal injury during neonatal aortic arch reconstruction. J Thorac Cardiovasc Surg 2012;144:1323–8–
[18]
AC C
1328.e1–2.
Polimenakos AC, Wojtyla P, Smith PJ, Rizzo V, Nater M, Zein El CF, et al. Postcardiotomy extracorporeal cardiopulmonary resuscitation in neonates with complex single ventricle: analysis of outcomes. Eur J Cardiothorac Surg 2011;40:1396–405–
discussion1405.
[19]
Ravishankar C, Dominguez TE, Kreutzer J, Wernovsky G, Marino BS, Godinez R, et al. Extracorporeal membrane oxygenation after stage I reconstruction for hypoplastic left heart syndrome. Pediatr Crit Care Med 2006;7:319–23.
[20]
Sherwin ED, Gauvreau K, Scheurer MA, Rycus PT, Salvin JW, Almodovar MC, et al.
ACCEPTED MANUSCRIPT
Extracorporeal membrane oxygenation after stage 1 palliation for hypoplastic left heart
AC C
EP
TE D
M AN U
SC
RI PT
syndrome. J Thorac Cardiovasc Surg 2012;144:1337–43.
S0022-5223(16)30086-1
DOI:
10.1016/j.jtcvs.2016.03.075
Reference:
YMTC 10503
To appear in:
The Journal of Thoracic and Cardiovascular Surgery
Received Date: 15 July 2015 Revised Date:
26 March 2016
Accepted Date: 30 March 2016
Please cite this article as: Botha P, Deshpande SR, Wolf M, Heard M, Alsoufi B, Kogon B, Kanter K, Extra-corporeal membrane oxygenator support in infants with systemic-pulmonary shunts, The Journal of Thoracic and Cardiovascular Surgery (2016), doi: 10.1016/j.jtcvs.2016.03.075. 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
Title Page Title: Extra-corporeal membrane oxygenator support in infants with systemic-pulmonary shunts. Running Head: ECMO in Infants with SP Shunts
RI PT
Authors: Phil Botha PhD, FRCS, Shriprasad R. Deshpande MBBS MS, Michael Wolf MD, Michael Heard RRT, Bahaaldin Alsoufi MD, Brian Kogon MD and Kirk Kanter MD . Institutions and Affiliations: Children’s Healthcare of Atlanta, Emory University, Atlanta, Georgia.
Corresponding Author:
EP
TE D
Shriprasad R. Deshpande Assistant Professor Department of Pediatrics Division of Pediatric Cardiology Emory University School of Medicine Children's Healthcare of Atlanta 1405 Clifton Road Atlanta, GA, USA 30322-1101 Email: [email protected] Phone: 404-694-7739 Fax : 770-488-9480
M AN U
Word Count: 3441 (manuscript word count)
SC
Keywords: Extracorporeal membrane oxygenation, ECMO; Shunts (Aorto-pulmonary); Congenital heart disease; CHD, Univentricular heart; Pediatric.
Disclosure : None of the authors have any conflicts of interest to disclose. Funding : No funding was provided or utilized for the conduct of this study
AC C
Abbreviations : ECMO : Extracorporeal membrane oxygenation SP shunt : systemic-pulmonary shunt BT shunt : modified Blalock- Taussig shunt eCPR : Extra-corporeal cardiopulmonary resuscitation CVVH : Continuous veno-venous hemofiltration
ACCEPTED MANUSCRIPT
Abstract Background
RI PT
Management of a patent systemic-pulmonary shunt and the resultant run-off whilst on extracorporeal membrane oxygenation (ECMO) varies amongst institutions. We have employed a strategy of increased flow without surgical reduction of the shunt diameter, and describe our results with this strategy. Methods
SC
Database review of 169 successive veno-arterial ECMO runs in infants and neonates. ECMO flow, time taken to achieve lactate clearance, normal pH and negative fluid balance was
M AN U
contrasted in patients with and without shunts. Results
Between 2002 and 2013, 51 of 169 (30.2%) infants had a shunt in situ when ECMO was initiated. Significantly higher ECMO flows were maintained in the shunt group (161±43 ml/kg/min vs. 134±41 ml/kg/min, P<0.001). Infants with shunts had significantly higher preECMO and peak lactates (12.4±5.6 mmol/l vs. 10.0±6.3 mmol/l, p<0.05, and 13.7±4.9 mmol/l
TE D
vs. 11.6±5.5 mmol/l, p<0.02 respectively) and clearance required a longer period of support (median 28.8, 16.1 – 63.3h vs. 17.5, 10.8 – 34.5h, p<0.001). Although the absolute rate of lactate clearance was not significantly different between groups (Median 0.46, 0.12 – 0.72 mmol/l/h vs. 0.48, 0.22 – 0.86mmol/l/h, p = 0.139) the presence of a shunt, neonatal age, peak lactate, extra-
EP
corporeal cardiopulmonary resuscitation and the use of hemofiltration on ECMO significantly predicted the rate of clearance. Survival to hospital discharge was similar in the shunt and non-
AC C
shunt groups (49.0% vs. 48.3%, p=0.932). Conclusions
A strategy of increased ECMO flow without surgically restricting shunt diameter appears successful in providing circulatory support in the majority of patients with systemic-pulmonary shunts. Equivalent survival suggests that routine surgical reduction of shunt diameter is not indicated.
Word Count: 249
ACCEPTED MANUSCRIPT
Introduction
2
The circulation in the presence of a surgical systemic-pulmonary (SP) shunt is one of inherent
3
instability, particularly in the setting of single ventricle physiology. Fluctuations in either the
4
systemic or pulmonary vascular tone can result in a profound and rapid alteration of systemic
5
perfusion. In the presence of systemic diastolic run-off into the low resistance pulmonary bed,
6
any reduction in the systemic perfusion pressure can lead to a rapid decompensation of
7
myocardial perfusion and impaired cardiac output in return. In addition to a tendency to periods
8
of pulmonary over-circulation, acute thrombosis of the shunt will similarly precipitate a
9
profound collapse of the cardiovascular system. For these reasons, the clinician caring for
SC
RI PT
1
patients with shunts is not infrequently faced with a need to support the collapsed circulation
11
using extra-corporeal membrane oxygenation (ECMO).
12
Initial reports of ECMO for post-cardiotomy circulatory support considered patients with
13
systemic-pulmonary shunts unsuitable. The resultant run-off, increased pulmonary flow and
14
therefore pulmonary venous return were thought to preclude successful support using this
15
modality[1,2]. Complete occlusion of the shunt by means of a surgical clip at the time of ECMO
16
was shown to lead to unacceptable mortality rates and many centers have therefore adopted a
17
strategy of increased flow to compensate for shunt run-off[3]. It has been suggested that partial
18
occlusion of the shunt during ECMO may reduce run-off into the pulmonary bed, resulting in
19
improved perfusion pressure and hasten resolution of the malperfusion state[4]. Our institution
20
has pursued a strategy of liberal flow on ECMO without applying any limitation to shunt run-off.
21
Here, we review our results using this strategy in terms of resolution of the malperfusion state
22
and other outcomes.
EP
TE D
M AN U
10
AC C
23 24
Methods
25
Our prospectively collected ECMO database was interrogated to identify all patients that
26
underwent ECMO for cardiac indications between January 2002 and October 2013. This data
27
was supplemented by chart review and examination of the electronic record of laboratory
28
investigations. Neonates and infants that underwent veno-arterial ECMO whilst having an
29
systemic-pulmonary shunt in situ (Shunt group) were contrasted to all infants and neonates
30
without shunts during the same period (Non-shunt group). Patients that underwent the first stage
31
Norwood procedure with a Sano RV-PA conduit were not included in the shunt group. In
ACCEPTED MANUSCRIPT
32
patients that underwent pre- and post-cardiotomy ECMO, only the post-operative episode was
33
analyzed.
34 Site of ECMO cannulation was dictated by the clinical situation. Cannulation was performed via
36
central cannulation (reopening of the sternotomy incision ) in the early post-operative period and
37
via a neck incision (jugular and carotid vessels ) in those requiring ECMO after the two-week
38
post-operative period.
39
With cannulation via the right carotid artery, we aim to place the tip of the arterial cannula at the
40
junction of the innominate artery and the aortic arch. This does produce the possibility of the
41
systemic end of the shunt being occluded by the cannula. We do however routinely continue
42
ventilation in this setting, and a loss of end-tidal CO2 would demonstrate that the shunt is being
43
occluded by the cannula. We have not observed this phenomenon. We further confirm the
44
patency with routine echocardiogram during the ECMO as well as during the weaning trials.
45
Again, if there is any concern about the cannula placement /shunt occlusion, the cannulae are
46
repositioned.
47
Similarly for patients with an open chest post sternotomy, direct central cannulation was
48
performed which involved cannulation of the right atrium (venous) and the aorta (arterial).
49
ECMO was delivered using a standard circuit (S-97-E Tygon tubing, Medtronic, Minneapolis,
50
MN), oxygenator (pediatric Quadrox D ©Maquet, Rastatt, Germany) and roller pump (Century,
51
Mesa, AZ) during this period. ECMO flows were increased to around 100ml/kg/min in non-
52
shunted patients and up to 200ml/kg/min in patients with systemic-pulmonary shunts. Flow rate
53
was adjusted according to pre-membrane oxygen saturation and blood pressure thereafter.
SC
M AN U
TE D
EP
AC C
54
RI PT
35
55
Serum lactic acid was measured using a regularly calibrated point-of care analyzer (iSTAT,
56
Abbott Point of care) and is presented as mmol/l. The level of serum lactate prior to
57
commencement of ECMO and the peak serum lactate measured during ECMO support was
58
documented. Typically, serum lactates were measured approximately two hourly in the early
59
phase after institution of ECMO and the frequency of sampling adjusted according to clinical
60
need thereafter. Time to lactate clearance was defined as the time elapsed between the institution
61
of ECMO and the serum lactate reaching a value below 2.0mmol/l (the upper limit of normal in
62
our laboratory). Rate of lactate clearance was calculated as peak lactate – 2.0, divided by the
ACCEPTED MANUSCRIPT
time elapsed between initiating ECMO and the first lactate level below 2.0mmol/l recorded. For
64
analysis of variables affecting the rate of lactate clearance, the range of clearance rates was
65
divided into equal quartiles and entered into univariate and multivariate logistic regression
66
models. For multivariate analysis, all variables with p values less than 0.10 in univariate analysis
67
were entered into the multinomial logistic regression model in a stepwise fashion. This analysis
68
considered the highest quartile of lactate clearance as the reference category and parameter
69
estimates for the contrast with the lowest quartile are reported.
70
Arterial blood gas analysis was performed at hourly intervals initially after institution of ECMO
71
using the same point-of-care analyzer as for lactic acid measurements. The fluid balance was
72
recorded in the electronic record throughout and analyzed in 12-hour intervals from 07:00 to
73
19:00 and 19:00 to 07:00. The time at conclusion of the first 12-hour interval after the initiation
74
of ECMO during which a negative fluid balance was achieved for that 12-hour period was taken
75
as the time at which a negative fluid balance was attained for the analysis. Total fluid balance
76
was calculated as total measured intake minus total measured fluid loss, not accounting for
77
insensible losses.
M AN U
SC
RI PT
63
78
Statistical analyses we're performed using the SPSS v.21. Parametric continuous variables are
80
presented as mean ± standard deviation and analyzed using the student’s t-test. Non-parametric
81
variables are presented as median followed by the 25th - 75th percentiles and analyzed using the
82
Wilcoxon rank-sum test or Mann-Whitney test where indicated. In all analyses, a p-value less
83
than 0.05 was considered statistically significant. Our institutional review board has reviewed the
84
study protocol and has waived the need for individual informed consent. All authors have had
85
full access to the data and have read and agreed to this final version of the manuscript.
87
EP
AC C
86
TE D
79
88
Results
89
During the period January 2002 to October 2013, 218 patients underwent 224 ECMO runs for
90
cardiac indications. Of these, 169 patients were neonates or infants at the time of commencing
91
ECMO ; 51 with an systemic-pulmonary shunt and the remaining 118 representing the non-shunt
92
group. One patient with a diagnosis of Pulmonary atresia/ ventricular septal defect with a
93
modified Blalock - Taussig (BT) shunt placed at a prior admission underwent veno-venous
ACCEPTED MANUSCRIPT
ECMO for a suspected occluded shunt and was therefore not included in the analysis.
95
Demographic variables in the shunt and non-shunt groups are contrasted in Table 1. Age was
96
not significantly different in the shunt and non-shunt groups (median 15 and 19 days
97
respectively, p=0.384), but infants with shunts weighed significantly less (3.3 vs. 3.6kg,
98
p<0.010). A greater proportion of shunt patients underwent ECMO during the immediate post-
99
operative admission (94.1% vs. 76.3%, p<0.005).
100
RI PT
94
ECMO conduct
102
ECMO was instituted in the operating room after failure to wean from cardiopulmonary bypass
103
or for marginal hemodynamics after weaning in 20 patients in the non-shunt group (16.9%) and 6
104
patients in the shunt group (11.9%, p=0.489). The remainder were cannulated in the cardiac
105
intensive care unit. Of the ones cannulated in the CICU, the common indication was cardiac
106
arrest. These patients were therefore resuscitated using the extra-corporeal cardiopulmonary
107
resuscitation (ECPR) pathway. A greater proportion of shunt patients underwent (ECPR)
108
protocol (72.5% vs. 47.5%, p<0.010). As shown in table 1, other indications included
109
cardiovascular instability, hypoxia, respiratory arrest and concern for shunt thrombosis. This
110
distribution was comparable in the shunt and non-shunt groups. Decision regarding
111
cardiovascular instability was usually based on increasing needs for inotropic and vasopressor
112
support, increase in lactic acidosis, as well as other measures of low cardiac output such as low
113
NIRS, poor perfusion, narrow pulse pressure with tachycardia. The decision was made by the
114
intensivist in conjunction with the primary surgeon. Patients who were cannulated for the
115
concern of shunt thrombosis (5/51) underwent a diagnostic cardiac catheterization /angiography
116
to visualize the shunt patency after stabilization. Shunt patency was established in each one of
117
these patients via angiography and or intervention as needed within 24 hours of initiation of
118
ECMO. All five of these patients therefore had a patent shunt through the rest of their ECMO
119
course.
M AN U
TE D
EP
AC C
120
SC
101
121
Cannulation was performed via the neck vessels in 10 patients in the shunt group (19.6%) and 33
122
patients in the non-shunt group (28.0%, p=0.336), and via the open chest in the remainder.
123
Significantly higher ECMO flows were maintained in the shunt group at 4 hours (median 161±43
ACCEPTED MANUSCRIPT
124
ml/kg/min vs. 134±41 ml/kg/min, p<0.001) and at 24 hours (median 148±43 ml/kg/min vs.
125
124±38 ml/kg/min, P<0.01).
126 Lactate Clearance
128
Serum lactic acid levels were significantly higher in patients with SP shunts immediately prior to
129
going onto ECMO than in patients without shunts (12.4±5.6 mmol/l vs. 10.0±6.3 mmol/l,
130
p<0.05). Similarly lactate levels peaked significantly higher in patients with shunts (13.7±4.9
131
mmol/l vs. 11.6±5.5 mmol/l, p<0.02). Five patients died or were de-cannulated before lactate
132
clearance could be achieved (3 in the shunt group and 2 in the non-shunt group). A significantly
133
greater proportion of patients in the shunt group required more than 48 hour to achieve lactate
134
clearance (31.4% vs. 12.7%, p=0.004). Median time to achieve lactate clearance was
135
significantly longer in patients with systemic-pulmonary shunts 28.8, 16.1 – 63.3h vs. 17.5, 10.8 –
136
34.5h, p<0.001 . Figure 1 demonstrates graphically the pre-ECMO and peak lactic acid levels, and
137
median time to lactate clearance in patients with and without shunts requiring ECMO. Although
138
the rate of lactate clearance appeared to be lower in shunt patients, the median difference did not
139
reach statistical significance (Median 0.46, 0.12 – 0.72 mmol/l/h vs. 0.48, 0.22 – 0.86mmol/l/h, p
140
= 0.139). Results of univariate and multivariate analysis for factors affecting lactate clearance are
141
listed in Table 2. In the multivariate model, peak serum lactic acid during ECMO, ECPR (as
142
opposed to elective institution of ECMO), neonatal age and the use of continuous veno-venous
143
hemofiltration (CVVH) were significant predictors of lactate clearance. The presence of a
144
systemic-pulmonary shunt was significantly predictive of belonging to the lowest quartile for
145
rate of lactate clearance but the interaction term size of shunt/body weight was not a statistically
146
significant predictor. There was no relationship between the ECMO flow rate and clearance of
147
lactate and this was independent of the presence or absence of SP shunt.
SC
M AN U
TE D
EP
AC C
148
RI PT
127
149
Acidemia/ Fluid balance
150
The median lowest arterial blood pH recorded in the period immediately prior to ECMO
151
cannulation or whilst on ECMO was 7.120, 7.022 – 7.225 in the shunt group and 7.183, 7.054 –
152
7.244 in the non-shunt group. This difference did not reach statistical significance on the Mann-
153
Whitney U-test (p=0.074). Time taken to achieve an arterial blood pH of greater than 7.350 on
154
ECMO also did not differ significantly between the shunt and non-shunt groups (median 0.62,
ACCEPTED MANUSCRIPT
0.33 – 1.07h vs. 0.68, 0.23 – 1.15h, p=0.822.) Data on fluid balance during ECMO was available
156
for 24 patients in the shunt group and 70 patients in the non-shunt group. Median time to achieve
157
a 12-hour period of overall negative fluid balance was 47.0, 34.6 – 59.2h vs. 31.9, 19.1 – 49.1h
158
in the shunt and non-shunt groups respectively (p=0.069). The overall fluid balance was not
159
significantly different in shunt and non-shunt groups, irrespective of normalization for the
160
duration of ECMO (+24.6±657.3 ml/day vs. +48.9±888.4 ml/day, p=0.385).
RI PT
155
161 Norwood Stage 1
163
A pre-specified subgroup analysis was performed using the cohort of patients that required
164
ECMO support in the post-operative period after stage 1 Norwood procedure. This comprised a
165
group of 44 patients, 16 (36%) having received modified BT shunts and 28 (64%) having
166
received Sano (right ventricle to pulmonary artery) conduits. The size of the BT shunt was 3mm
167
in 3 patients and 3.5mm in 14 patients. Table 3 summarizes demographic and ECMO variables
168
contrasted in these two groups. Patients in these two groups had similar mean weights and a
169
similar proportion underwent ECMO cannulation as part of an ECPR protocol. Pre-ECMO
170
lactates and peak lactates were not significantly different between the two groups. Norwood
171
stage 1 patients with BT shunts had a lower rate of lactate clearance on ECMO than patients with
172
Sano conduits (0.29, 0.08 – 0.56 vs. 0.47, 0.28 – 0.82 mmol/l/hr, Mann-Whitney p=0.052).
173
TE D
M AN U
SC
162
Outcomes
175
Overall median duration of ECMO for infants and neonates was 119.3 hours (Range 3.3 - 651.7
176
hours). Duration of ECMO support was significantly greater in the non-shunt group (131.0, 78.4
177
- 195h vs. 77.6, 41.7 - 175h, p=0.021). Surrogate measures of low cardiac output or malperfusion
178
state such as renal insufficiency were also assessed as secondary endpoints. In addition to the
179
fluid balance mentioned above (which reflected the urine output), serum creatinine as well as
180
need for CVVH post-ECMO was compared. serum creatinine values at decannulation or at the
181
time of discharge. Only two patients needed CVVH post-ECMO , both were on CVVH during
182
the ECMO run. Both these patients belonged to the shunt group. Due to a very small number
183
(2/169 patients), further analysis related to persistent renal function post- ECMO , was not
184
performed.
AC C
EP
174
ACCEPTED MANUSCRIPT
For the entire cohort of patients receiving ECMO support over this time period, survival to
186
ECMO de-cannulation was 71.1% and survival to hospital discharge was 50.9%. Survival for
187
infants and neonates overall was 48.5%, and not significantly different in shunt and non-shunt
188
groups (49.0% vs. 48.3%, p=0.932). In the cohort of patients post stage 1 Norwood, survival was
189
nearly identical, being 71.1% survival to ECMO de-cannulation and 46.7% survival to discharge.
RI PT
185
190 Discussion
193
Palliation of single ventricle patients with a staged approach is now the standard accepted
194
strategy [5]. In conditions with inadequate pulmonary blood flow and those requiring single
195
ventricle palliation, the systemic-pulmonary shunt still represents in many patients the first
196
surgical step (or part thereof). The period from the creation of the shunt until definitive repair, or
197
the substitution of the systemic-pulmonary shunt with a cavo-pulmonary shunt, remains the most
198
critical phase in terms of risk for mortality and morbidity [6]. The shunt-dependent circulation is
199
one characterized by lability, as the synthetic conduit used for the majority of such shunts does
200
not allow for the usual degree of cardiovascular auto-regulation. Creating a connection between
201
the usually serial pulmonary and systemic circulations can lead to rapid and profound swings in
202
perfusion in favor of either vascular bed. Patient populations undergoing SP shunts have changed
203
considerably over the past 70 years[7,8] but the rate of in-hospital mortality has changed little
204
from 8.1% in the earliest reports[9] to rates of around 4-10% in the higher risk contemporary
205
populations [8,10,11]. Major morbidity also remains common after SP shunts, with re-
206
intervention required in 7.3% and around 3-6% requiring ECMO support[11,12]. We have
207
analyzed a consecutive series of 169 infants and neonates requiring mechanical circulatory
208
support in the form of veno-arterial ECMO, to ascertain the impact of the presence of an
209
systemic- pulmonary shunt during ECMO support on their outcomes.
M AN U
TE D
EP
AC C
210
SC
191 192
211
In early series of pediatric post-cardiotomy ECMO, patients with shunts were considered not to
212
be suitable candidates because of the limitations shunt run-off would impose on systemic
213
perfusion [1,2]. Jaggers and colleagues reported outcomes in 9 patients with shunts requiring
214
ECMO [3]. The authors occluded the shunt surgically at the time of going on to ECMO in 4
215
patients, but had no survivors in this group. In contrast, the mortality rate in 5 patients where the
ACCEPTED MANUSCRIPT
shunt was left patent and the ECMO flow increased to compensate for shunt run-off was
217
considerably lower at 20%. More recently, Allan and colleagues reported 44 patients that
218
underwent ECMO following single ventricle palliation with a systemic to pulmonary arterial
219
shunt [4]. The authors maintained shunt patency wherever possible, but in patients demonstrating
220
decreased systemic perfusion with delayed clearance of lactate despite increasing ECMO flow, 1
221
or more surgical clips were placed to restrict the shunt diameter and reduce run-off. This was
222
required in 17 of 44 patients (39%). Overall survival to hospital discharge was 48%, with 20% of
223
survivors requiring shunt clipping, as opposed to 61% of non-survivors. In the present series we
224
have found shunt patients requiring ECMO to have lower body weights, to have required ECMO
225
predominantly in the post-operative period and more commonly as part of extra-corporeal
226
resuscitation for cardiovascular collapse. Significantly higher ECMO flows were maintained in
227
the shunt group and no surgical reduction of shunt diameter was undertaken. Survival to
228
discharge in all infants and neonates requiring ECMO in our series was comparable to these
229
previous publications and not affected by the presence of a systemic-pulmonary shunt.
M AN U
SC
RI PT
216
230
As a surrogate measure of tissue oxygen delivery and anaerobic metabolism, the serum lactic
232
acid at various post-operative time-points has been shown to predict outcomes in pediatric
233
cardiac surgery [13-15]. Several factors will determine the level of serum lactate in an infant
234
requiring ECMO. The duration and severity of the low cardiac output/ hypoxemic state
235
prompting the need for ECMO will be major determinants, but other factors including duration
236
of cardiopulmonary bypass, circulatory arrest and/ or regional perfusion have also been shown to
237
correlate with early post-operative serum lactate [16,17]. As patients in the shunt group in our
238
series were usually in the peri-operative period and more likely to have required ECMO after
239
cardiac arrest and therefore as part of an eCPR protocol, we found not surprisingly that pre-
240
ECMO serum lactic acid was significantly higher in this group. The determinants of the rate of
241
lactate clearance in a patient on ECMO are also multiple. Whether the precipitant of
242
cardiovascular collapse has resolved will have a profound effect, as will the presence of residual
243
uncorrected anatomical abnormalities. The adequacy of organ perfusion once on ECMO will
244
determine the level of ongoing lactate production and the rate at which the existing pool will be
245
metabolized. Although the rate of lactate clearance in shunted patients in our series appeared to
246
be lower, the median difference in the rate of clearance on ECMO did not reach statistical
AC C
EP
TE D
231
ACCEPTED MANUSCRIPT
significance. On multivariate analysis peak serum lactic acid during ECMO, ECPR, neonatal age
248
and the use of continuous veno-venous hemofiltration (CVVH) were significant predictors of the
249
rate of lactate clearance. The presence of a systemic-pulmonary shunt was also a significant
250
predictor of belonging to the lowest quartile for rate of lactate clearance but the relative size of
251
the shunt did not appear significant. It has been a concern that achieving and maintaining higher
252
ECMO flows in patients with a patent systemic-pulmonary shunt may be at the expense of a
253
greater need for volume administration, and therefore delay the resolution of fluid sequestration.
254
Median time in hours to achieve a 12-hour period of overall negative fluid balance was 47.0,
255
34.6 – 59.2h vs. 31.9, 19.1 – 49.1h in the shunt and non-shunt groups respectively. Although
256
many factors could have influenced this trend, including lower renal perfusion and a higher
257
requirement for CVVH in the shunt group, the difference did not reach statistical significance
258
(p=0.069). The lower number of patients in which all the required datapoints were available,
259
cautions the possibility of a type-2 error in this analysis.
M AN U
SC
RI PT
247
260
We also analyzed as a more homogenous subgroup, all patients that required ECMO after stage 1
262
Norwood contrasting those with systemic-pulmonary shunts with those with RV-PA (Sano)
263
shunts. Although pre-ECMO and peak lactates were not significantly different between groups,
264
patients with BT shunts had a significantly lower rate of lactate clearance than patients with Sano
265
conduits. Polimenakos and colleagues reported on 20 functional single ventricle patients
266
supported with ECMO post cardiotomy[18]. They left the shunt patent in all cases in the hope of
267
minimizing pulmonary vascular endothelial injury and utilized higher pump flows
268
(>150ml/kg/min) to compensate for shunt runoff. Vaso-active-inotropic support was continued
269
and 20ppm Inhaled Nitric oxide administered in all cases. Twelve of these patients had
270
pulmonary blood flow established via a Sano shunt, which would likely have significantly
271
smaller impact on flow requirement. Survival to discharge was 57%. Ravishankar and colleagues
272
also document a survival to discharge of 39 % in 36 neonates rescued by ECMO following stage
273
1 palliation[19]. They note that the shunt was left patent (modified BT shunt in the majority of
274
cases), requiring flows between 150-200ml/kg/min to compensate for shunt run-off. Similarly,
275
the analysis from the ELSO registry [20] showed that for patients HLHS after stage I palliation,
276
needing ECMO, 59% survived the decannulation while 31% survived to discharge. The risk
277
factors associated with death were Black race, lower weight and longer duration of ventilator
AC C
EP
TE D
261
ACCEPTED MANUSCRIPT
support prior to ECMO. However, they were unable to comment on the type of shunt (SP shunt
279
versus a Sano), other surgical details (such as clipped shunts) or patient acuity prior to ECMO. In
280
our experience, despite a slower clearance of lactate in patients with systemic-pulmonary shunts,
281
survival to discharge was not affected irrespective of the shunt type and was uniformly better
282
(48%) than the ELSO registry report.
RI PT
278
283 284 Limitations
286
The study is inherently limited in generalizability by its retrospective nature and sample size.
287
Although a standardized protocol for the management of all patients on ECMO was followed,
288
this will have undergone some evolution over the 11year time-course of the study.
M AN U
SC
285
289 Conclusions
291
Mortality remains common in Infants requiring support by Extra-corporeal membrane
292
oxygenation. Accepting the limitations of mortality as an endpoint, the presence of a systemic-
293
pulmonary shunt during ECMO support does not appear to have an adverse effect on outcome, if
294
ECMO flow is appropriately increased to compensate for shunt run-off. Although patients with a
295
patent systemic-pulmonary shunt appear to have a delay in the resolution of the malperfusion
296
state on multivariate analysis, routine surgical reduction of shunt diameter to reduce shunt run-
297
off cannot be supported based on our data.
AC C
EP
TE D
290
ACCEPTED MANUSCRIPT
Figure Legend Non-shunt group (N=118)
p
Age (days)
15 (8-38)
19 (7-109)
N.S.
Weight (kg)
3.3 (2.7-3.8)
3.6 (2.9 - 5.4)
0.010
Gender: Male
69 (58.5%)
27 (52.9%)
N.S.
Diagnosis: HLHS
10 (19.6%)
23 (19.5%)
PA/IVS
11 (21.6%)
4 (3.4%)
Other Single ventricle
22 (43.1%)
TAPVR
0
DCM/ Myocarditis
0
Truncus arteriosus Other Post-operative
ECPR Shunt diameter: 3.0mm 3.5mm
SC 11(9.3%) 9 (7.6%)
0
8 (6.8%)
7 (13.7%)
37 (31.4%)
48 (94.1%)
90 (76.3%)
0.005
16 (31.4%)
28 (23.7%)
N.S.
37 (72.5%)
56 (47.5%)
0.003
8 (15.7%) 41 (80.4%) 2 (3.9%)
EP
4.0mm
15 (12.7%)
1 (2.0%)
TE D
Norwood stage 1
11 (9.3%)
M AN U
TGA
RI PT
Shunt Group (N=51)
Indication for ECMO: Unable to wean CPB
13 (11.01%)
Cardiac Arrest
19 (37.3%)
37 (31.4%)
CV Instability
16 (31.4%)
38+7 (38.1%)
Hypoxia
6 (11.8%)
11 (9.3%)
Respiratory arrest
7 (6.1%)
3 (2.5%)
AC C
2 (3.9%)
Concern for shunt 5 (9.8%) Thrombosis
0
Table 1. Patient demographics. The shunt group comprises neonatal and infant patients requiring VAECMO support during the study period. N.S. - Not significant, HLHS- Hypoplastic left heart syndrome, PA/IVS - Pulmonary atresia with intact ventricular septum, TAPVR - Total anomalous pulmonary venous return, DCM- Dilated cardiomyopathy, TGA, Transposition of the great arteries, ECPR - Extracorporeal cardiopulmonary resuscitation, CPB - Cardiopulmonary bypass, CV – cardiovascular.
Multivariate Coefficient
95% CI
Neonatal age
0.013
0.029
4.291
1.160-15.866
Weight (kg)
0.111
ECPR
<0.001
SP Shunt
0.174
0.019
4.926
Shunt diameter / weight (mm/kg)
0.409
Pre-ECMO lactate
0.012
Peak lactate
<0.001
<0.001
CVVH
0.006
0.050
Stage 1 Norwood
0.344
CPB time (min)
0.148
Crossclamp (min)
0.968
Major bleeding
0.158
Neurological event
0.595
SC
RI PT
Multivariate p
TE D
Univariate p
M AN U
ACCEPTED MANUSCRIPT
1.303-18.681
0.729
0.634-0.839
0.280
0.079-0.999
AC C
EP
Table 2. Results of univariate and multivariate analyses for factors predictive of delayed lactate clearance. ECPR – Extracorporeal cardiopulmonary resuscitation, SP – Systemic-pulmonary, CVVH – continuous veno-venous hemofiltration, CPB- cardiopulmonary bypass.
ACCEPTED MANUSCRIPT
Sano group (N=28)
p
Weight (kg)
3.28 ±0.67
3.11 ±0.67
0.493
ECPR
9 (56%)
17 (61%)
0.954
Indication for ECMO: Unable to wean CPB
2 (13%)
2 (7%)
Cardiac Arrest
2 (13%)
12 (43%)
CV instability
5 (31%)
11 (39%)
Hypoxia
1 (6%)
2 (7%)
Respiratory arrest
0
Metabolic
2 (13%)
Pre ECMO Lactate (mmol/l) Peak Lactate (mmol/l) Rate of Lactate clearance (mmol/l/hr) Time to pH > 7.35 (hr)
Total fluid balance (ml/day)
SC 0
4 (25%)
0
13.0 (5.2)
11.4 (5.1)
0.362
13.2 (4.4)
13.1 (4.9)
0.407
0.29 (0.32)
0.47 (0.41)
0.052
0.63 (0.46)
0.71 (2.36)
0.464
45.7 (20.5)
28.6 (20.95)
0.185
2.0 (310.0)
0.129
TE D
Time to negative fluid balance (hours)
1 (4%)
M AN U
Shunt Thrombosis
RI PT
BT Shunt group (N=16)
-116.1 (430.9)
AC C
EP
Table 3. Comparison of ECMO conduct and metabolic normalization in first stage Norwoood patients undergoing ECMO, contrasted in the subset with systemic-pulmonary shunts and those with right ventricle to pulmonary artery conduits.
ACCEPTED MANUSCRIPT
20 18
*
*
Shunt No Shunt
16
RI PT
* - p <0.05
12
Rate of lactate clearance: (Median 0.46, 0.12 – 0.72 mmol/l/h vs. 0.48, 0.22 – 0.86 mmol/l/h, p = 0.139)
10
SC
Lactate (mmol/l)
14
8
4
* 2 0 0
5
10
15
20
M AN U
6
25
30
35
40
45
50
55
60
65
TE D
Time (h)
AC C
EP
Figure 1. Line plot depicting mean lactic acid levels immediately prior to ECMO support and peak lactate level on ECMO, as well as median time to achieve normal lactic acid level in systemic-pulmonary shunt and non-shunt groups. Error bars depict standard deviation for parametric variables and 25-75% interquartile range for time to normalization of lactate.
ACCEPTED MANUSCRIPT
References [1]
Kanter KR, Pennington G, Weber TR, Zambie MA, Braun P, Martychenko V.
RI PT
Extracorporeal membrane oxygenation for postoperative cardiac support in children. J Thorac Cardiovasc Surg 1987;93:27–35. [2]
del Nido PJ, Dalton HJ, Thompson AE, Siewers RD. Extracorporeal membrane
oxygenator rescue in children during cardiac arrest after cardiac surgery. Circulation
[3]
SC
1992;86:II300–4.
Jaggers JJ, Forbess JM, Shah AS, Meliones JN, Kirshbom PM, Miller CE, et al. Extracorporeal membrane oxygenation for infant postcardiotomy support: significance
[4]
M AN U
of shunt management. Ann Thorac Surg 2000;69:1476–83.
Allan CK, Thiagarajan RR, del Nido PJ, Roth SJ, Almodovar MC, Laussen PC. Indication for initiation of mechanical circulatory support impacts survival of infants with shunted single-ventricle circulation supported with extracorporeal membrane oxygenation. J Thorac Cardiovasc Surg 2007;133:660–7.
[5]
Williams JA, Bansal AK, Kim BJ, Nwakanma LU, Patel ND, Seth AK, et al. Two
TE D
thousand Blalock-Taussig shunts: a six-decade experience. Ann Thorac Surg 2007;84:2070–5–discussion2070–5. [6]
Fenton KN, Siewers RD, Rebovich B, Pigula FA. Interim mortality in infants with
7.
EP
systemic-to-pulmonary artery shunts. Ann Thorac Surg 2003;76:152–6–discussion156–
Blalock A. The Surgical Treatment of Malformations of the Heart. Jama 1945;128:189.
[8]
Heidari-Bateni G, Norouzi S, Hall M, Brar A, Eghtesady P. Defining the best practice
AC C
[7]
patterns for the neonatal systemic-to-pulmonary artery shunt procedure. J Thorac
Cardiovasc Surg 2014;147:869–873.e3.
[9]
de Leval MR, McKay R, Jones M, Stark J, Macartney FJ. Modified Blalock-Taussig shunt. Use of subclavian artery orifice as flow regulator in prosthetic systemic-
pulmonary artery shunts. J Thorac Cardiovasc Surg 1981;81:112–9. [10]
Myers JW, Ghanayem NS, Cao Y, Simpson P, Trapp K, Mitchell ME, et al. Outcomes of systemic to pulmonary artery shunts in patients weighing less than 3 kg: Analysis of shunt type, size, and surgical approach. J Thorac Cardiovasc Surg 2014;147:672–7.
ACCEPTED MANUSCRIPT
[11]
Petrucci O, O'Brien SM, Jacobs ML, Jacobs JP, Manning PB, Eghtesady P. Risk Factors for Mortality and Morbidity After the Neonatal Blalock-Taussig Shunt Procedure. Ann Thorac Surg 2011;92:642–52.
[12]
Guzzetta NA, Foster GS, Mruthinti N, Kilgore PD, Miller BE, Kanter KR. In-Hospital
RI PT
Shunt Occlusion in Infants Undergoing a Modified Blalock-Taussig Shunt. Ann Thorac Surg 2013;96:176–82. [13]
Siegel LB, Dalton HJ, Hertzog JH, Hopkins RA, Hannan RL, Hauser GJ. Initial
Intensive Care Med 1996;22:1418–23. [14]
SC
postoperative serum lactate levels predict survival in children after open heart surgery.
Charpie JR, Dekeon MK, Goldberg CS, Mosca RS, Bove EL, Kulik TJ. Serial blood
M AN U
lactate measurements predict early outcome after neonatal repair or palliation for complex congenital heart disease. J Thorac Cardiovasc Surg 2000;120:73–80. [15]
Schumacher KR, Reichel RA, Vlasic JR, Yu S, Donohue J, Gajarski RJ, et al. Rate of increase in serum lactate level risk-stratifies infants after surgery for congenital heart disease. J Thorac Cardiovasc Surg 2013.
[16]
Munoz R, Laussen PC, Palacio G, Zienko L, Piercey G, Wessel DL. Changes in whole
TE D
blood lactate levels during cardiopulmonary bypass for surgery for congenital cardiac disease: an early indicator of morbidity and mortality. J Thorac Cardiovasc Surg 2000;119:155–62. [17]
Algra SO, Schouten ANJ, van Oeveren W, van der Tweel I, Schoof PH, Jansen NJG, et
EP
al. Low-flow antegrade cerebral perfusion attenuates early renal and intestinal injury during neonatal aortic arch reconstruction. J Thorac Cardiovasc Surg 2012;144:1323–8–
[18]
AC C
1328.e1–2.
Polimenakos AC, Wojtyla P, Smith PJ, Rizzo V, Nater M, Zein El CF, et al. Postcardiotomy extracorporeal cardiopulmonary resuscitation in neonates with complex single ventricle: analysis of outcomes. Eur J Cardiothorac Surg 2011;40:1396–405–
discussion1405.
[19]
Ravishankar C, Dominguez TE, Kreutzer J, Wernovsky G, Marino BS, Godinez R, et al. Extracorporeal membrane oxygenation after stage I reconstruction for hypoplastic left heart syndrome. Pediatr Crit Care Med 2006;7:319–23.
[20]
Sherwin ED, Gauvreau K, Scheurer MA, Rycus PT, Salvin JW, Almodovar MC, et al.
ACCEPTED MANUSCRIPT
Extracorporeal membrane oxygenation after stage 1 palliation for hypoplastic left heart
AC C
EP
TE D
M AN U
SC
RI PT
syndrome. J Thorac Cardiovasc Surg 2012;144:1337–43.