Cardiopulmonary Bypass Increases Plasma Glial Fibrillary Acidic Protein Only in First Stage Palliation of Hypoplastic Left Heart Syndrome

Canadian Journal of Cardiology 32 (2016) 355e361

Clinical Research

Cardiopulmonary Bypass Increases Plasma Glial Fibrillary Acidic Protein Only in First Stage Palliation of Hypoplastic Left Heart Syndrome Luca Vedovelli, PhD,a Massimo Padalino, MD, PhD,b Manuela Simonato, PhD,a Sara D’Aronco, PharmD,c Diana Bertini, MD,b Giovanni Stellin, MD,b Carlo Ori, MD,d Virgilio P. Carnielli, MD, PhD,e and Paola E. Cogo, MD, PhDf a b

GPYIPR and PCare Laboratories, Pediatric Research Institute “Città della Speranza,” Padova, Italy

Pediatric and Congenital Cardiovascular Surgery Unit, Centro V. Gallucci, Padova University Hospital, Padova, Italy c d

e

Department of Women’s and Children’s Health, Padova University Hospital, Padova, Italy

Department of Medicine DIMED, Anesthesia and Resuscitation Institute, Padova University Hospital, Padova, Italy

Division of Neonatology, Department of Clinical Sciences, Polytechnic University of Marche and Azienda Ospedaliero-Universitaria Ospedali Riuniti, Ancona, Italy f

Pediatric Cardiac Anesthesia/Intensive Care Unit, Department of Pediatric Cardiology and Cardiac Surgery, Bambino Gesù Children’s Hospital, Rome, Italy

ABSTRACT

  RESUM E

Background: Univentricular congenital heart defects require openheart surgery soon after birth, and are associated with risk of brain injury and poor neurologic outcome. Methods: This is a prospective, observational study on children undergoing cardiac surgery. Plasma glial fibrillary acidic protein (GFAP), as an early marker of brain injury, was measured by ELISA at the end of anaesthesia induction, initiation of cardiopulmonary bypass (CPB), the end of cooling, the end of rewarming, the end of CPB, and after protamine administration. We recorded clinical and surgical parameters to assess which CPB phase and clinical parameters were associated with a GFAP increase. Results: We studied 13 children less than 50 months of age: 8 underwent Norwood or Damus-Kaye-Stansel palliation (group 1) and 5 underwent Fontan procedure (group 2). A GFAP increase was only observed in group 1, with the highest median value at the end of rewarming. No quantifiable levels of GFAP were measured at prebypass and the start of CPB stages in all patients. End of cooling

nitales uniIntroduction : Les malformations cardiaques conge cessitent une intervention chirurgicale à cœur ouvert ventriculaires ne es à un risque de le sions peu après la naissance, et sont associe re brales et à un mauvais pronostic neurologique. ce thodes : Ceci est une e tude observationnelle prospective sur les Me ine acide fibrillaire enfants subissant une chirurgie cardiaque. La prote coce d’une le sion ce re brale, a gliale (PAFG), en tant que marqueur pre te  mesure e par ELISA à la fin de l’induction d’une anesthe sie, à e l’initiation de la circulation extracorporelle (CEC), à la fin du refroichauffement, à la fin de la CEC et après dissement, à la fin du re  les paramètres l’administration de protamine. Nous avons enregistre valuer quelle phase de la CEC et quels cliniques et chirurgicaux pour e taient associe s avec une augmentation de la paramètres cliniques e PAFG. sultats : Nous avons e tudie  le cas de 13 enfants âge s de moins de Re 50 mois : 8 ont subi une chirurgie palliative de Norwood ou Damus-

Congenital heart defects (CHDs) affect nearly 1% of all births, and approximately 25% of patients have a critical CHD that needs a surgical intervention in the first year of life.1,2 Cardiac anomalies with univentricular heart physiology, such as hypoplastic left heart syndrome, are among the

most critical CHDs, and early surgery is the only way to grant survival, restoring blood oxygenation and hemodynamic balance between pulmonary and systemic flows.3 Improved surgical techniques4 and postoperative care in the last few decades have increased survival after surgery for CHDs, with an overall estimate of 3% early mortality.5,6 In this scenario, neurologic complications have emerged as a key issue in the long-term quality of life of children undergoing CHD surgery during infancy. Nearly 50% of children who had undergone surgery for a complex CHD requiring aortic cross-clamping or deep hypothermic circulatory arrest (DHCA) may show neurocognitive deficits at the

Received for publication March 25, 2015. Accepted June 24, 2015. Corresponding author: Dr Luca Vedovelli, Pediatric Research Institute “Città della Speranza, ” Corso Stati Uniti 4, 35127 Padova, Italy. Tel.: þ390498211477; fax: þ39-0499640146. E-mail: [email protected] See page 360 for disclosure information.

http://dx.doi.org/10.1016/j.cjca.2015.06.023 0828-282X/Ó 2016 Canadian Cardiovascular Society. Published by Elsevier Inc. All rights reserved.

356

Canadian Journal of Cardiology Volume 32 2016

and CPB-end GFAP, GFAP maximum value, and GFAP area under the curve all correlated with the CPB time spent at a cerebral regional saturation < 45% (P ¼ 0.021, 0.028, 0.007, 0.021, respectively). Conclusions: Children with univentricular heart defects exhibit a CPB plasma-GFAP increase only after stage 1 palliation. The maximum GFAP increase occurred at the end of rewarming. Further studies are needed to identify which clinical or surgical parameter(s) could reflect a GFAP increase during surgery for congenital heart defects, and whether GFAP levels correlate with the neurologic outcome.

Kaye-Stansel (groupe 1) et 5 ont subi une intervention de Fontan  te  observe e uniquement (groupe 2). Une augmentation de la PAFG a e diane la plus e leve e lors de la fin dans le groupe 1, avec une valeur me chauffement. Aucun niveau quantifiable de PAFG n’a e  te  mesure  du re rivation et au de but des e tapes de la CEC pour tous les paavant de tients. La PAFG en fin de refroidissement et en fin de CEC, la valeur maximale de PAFG, et l’aire de PAFG sous la courbe sont toutes en lation avec le temps de CEC passe  à une saturation ce re brale corre gionale <45 % (P ¼ 0,021, 0,028, 0,007, 0,021, respectivement). re Conclusions : Les enfants atteints de malformations cardiaques unisentent une augmentation de la PAFG plasmatique ventriculaires pre tape 1 de la chirurgie palliative. lors de la CEC seulement après l’e L’augmentation maximale de la PAFG a eu lieu à la fin du chauffement. D’autres e tudes sont ne cessaires pour de terminer re ter une quel(s) paramètre(s) clinique ou chirurgical pourrait refle augmentation de la PAFG pendant la chirurgie pour des malformations nitales, et si les niveaux de PAFG corrèlent avec cardiaques conge volution neurologique. l’e

beginning of school age.7 Several factors may contribute to these deficits, including prenatal lesions or technical issues during CHD surgery.8 Cerebral injury biomarkers8-14 seem to be a promising tool to identify ongoing brain damage during or closely after surgery and thus may effectively act to prevent further exacerbation of central nervous system insults. The glial fibrillary acidic protein (GFAP) is the principal intermediate filament in mature astrocytes. GFAP seems to fully fit the requirements of specificity, readiness of release, and ease of assaying required by a diagnostic brain injury neuromarker; in fact, GFAP can be measured in the peripheral blood and quantified proportionally to the degree of the injury.11,15 In this study, we evaluated GFAP plasmatic level variations during the different phases of cardiopulmonary bypass (CPB) in a small selected group of patients with univentricular congenital heart defects undergoing surgery. The GFAP rise was correlated with time-correspondent clinical parameters to asses if readily measurable biological changes or specific surgery phases could be related to the increase of the GFAP plasmatic level.

chromosomal abnormalities (ie, Down or Di George syndromes). According to the surgical procedure performed, patients could undergo aortic cross-clamp, DHCA, or selective regional cerebral perfusion. Neurologic risk time interval (NRTI) was defined as the period of time in which the patient was exposed at high risk of neurologic injury, that is, the time spent by the patient in selective regional cerebral perfusion plus the time of DHCA, whenever performed. Patients were divided into 2 groups: patients who underwent an NRTI during surgery (group 1) and patients who did not require an NRTI during surgery (group 2).

Patients and Methods Patients This is a prospective, observational clinical study in children with univentricular congenital heart defects undergoing different stages of palliation. The study was approved by the local Institutional Review Board and Ethic Committee (Padova University Hospital). Inclusion criteria were as follows: children with complex univentricular heart defects; elective cardiac surgery (patient on spontaneous breath before cardiac surgery); CPB time > 60 minutes (with aortic crossclamp > 20 minutes, when performed); hypothermia during CPB; stable hemodynamic conditions with constant inotropic support and constant volume loading for at least 3 hours before the study; and written informed consent. Exclusion criteria were as follows: age > 5 years; liver damage, factor V < 20% before surgery; kidney failure, with creatinine clearance < 30% before surgery; and preoperative diagnosis of

Surgery and samples collection After the induction of anaesthesia (fentanyl 5 mg/kg, thiopental 3 mg/kg or midazolam 0.2 mg/kg, and vecuronium bromide 0.1 mg/kg), infants were intubated and a central venous catheter was placed. General anaesthesia was obtained with fentanyl 50 mg/kg, cisatracurium besylate 3 mg/kg, and midazolam 3 mg/kg, infused at 1 mL/h or 2 mL/h, based on patient body weight (less or greater than 5 kg, respectively). After heparin 300 U/kg administration (activated clotting time [ACT] target ¼ 480 s), arterial and venous cannulation were performed and CPB was initiated. Hematic prime was used to maintain hematocrit between 25% and 30%. Temperature was monitored with nasopharyngeal and rectal probes. Hypothermia during CPB was defined as mild (35 C-30.1 C), moderate (30 C-25.1 C), and deep (25 C15.1 C).16 Whenever required, hematic cardioplegia, aortic cross-clamp, and DHCA or selective regional cerebral perfusion were used; adequate CPB flows were calculated based on body surface area, cardiac index, and minimal body temperature reached. Surgical procedures were performed according to univentricular heart defects palliation techniques (Norwood-like palliation, bidirectional cavopulmonary anastomosis, or Fontan completion-extracardiac conduit). Mixed venous saturation (SvO2) was recorded every 5 minutes with CDI Blood Parameter Monitoring System 500 (Terumo, Tokyo, Japan). Transcranial cerebral regional oxygen saturation (rO2) was measured with near infrared

Vedovelli et al. GFAP in Univentricular CHD

spectroscopy (NIRS) technology (INVOS, Somanetics, Troy, MI) every minute. The NIRS probe placement was unilateral. Blood gas analysis and metabolic parameters were measured at least every 20 minutes. At the end of surgery, all patients underwent rewarming to 37 C. Protamine 3 mg/kg (1 mg every 100 U of heparin) was administered at the end of CPB, for heparin reversal. Modified ultrafiltration was used in all patients. Blood samples for laboratory analysis were collected in EDTA-containing tubes (1.5 mL) from the central venous catheter before initiation of CPB and after protamine administration. In addition, samples were also collected from superior vena cava during CPB phases (CPB start, end of cooling, end of rewarming, and end of CPB before modified ultrafiltration). The choice of superior vena cava was warranted to confirm that the source of the collected blood was directly drained from the cerebral region. Samples analysis Blood tubes were centrifuged at 1400  g for 10 minutes to obtain plasma. Plasma was divided in 150 mL aliquots and stored at 80 C until analysis. GFAP was measured with a commercial ELISA kit RD192072200R (BioVendor, Brno, Czech Republic). The kit detection range was 0.25-25 ng/mL with a limit of detection (analyte giving absorbance higher than mean absorbance of blank) of 0.045 ng/mL. Nondetectable samples values were recorded as half of the limit of detection (0.0225 ng/mL). GFAP concentration was assessed, in duplicate, in all plasma samples. Calculations GFAP area under the curve (AUC) was automatically calculated with Prism 5.0 (GraphPad Software, La Jolla, CA). The cerebral regional saturation difference during CPB was calculated as the maximum rO2 value recorded by the INVOS device minus the minimum recorded value. GFAP was expressed as concentration (ng/mL of plasma) at each study point, as AUC during CPB time, and as the maximum GFAP concentration (GFAP max) reached by each patient. Statistical analysis Normal distribution was assessed by Shapiro-Wilk and Kolmogorov-Smirnov tests. Relationships between the surgical procedures, clinical parameters, and GFAP concentrations were assessed by a linear model of bivariate correlation and tested with Spearman’s r. Group comparison (group 1 vs group 2) was assessed by the Mann-Whitney U test. Statistical significance was set at P < 0.05. PASW Statistics 21.0 (IBM Corp., Armonk, NY) was used for the analysis. Results Patients Thirteen children with univentricular congenital heart defects were selected and studied in the Pediatric Cardiovascular Surgery Unit, Centro “V. Gallucci,” Padova University Hospital, Italy. Group 1 (with NRTI) included 8 patients who underwent first-time surgery with aortic cross-clamp; among them 7 required regional cerebral perfusion and 3

357

required DHCA. Six patients underwent Norwood operation, whereas Damus-Kaye-Stansel procedure was performed in 2. Group 2 (no NRTI) included 5 patients who underwent extracardiac Fontan procedure, with no aortic cross-clamp and no NRTI. CPB was conducted on alpha-stat strategy in all patients. Clinical and surgical characteristics are shown in Table 1. GFAP analysis No quantifiable GFAP plasma levels were measured in the pre-bypass and start of CPB samples in both groups. During and after CPB, GFAP increased only in group 1 (Table 2), and the highest median GFAP concentration was measured at the end of the rewarming phase. On the contrary, GFAP was not quantifiable in patients undergoing the Fontan procedure (group 2) without aortic cross-clamp and NRTI. The highest GFAP concentration was detected at the end of cooling in a Damus-Kaye-Stensel procedure with the longest aortic cross-clamp time (Fig. 1). In addition, in group 1, we also tested if there were significant correlations between the GFAP plasma levels and arterial lactates, pO2, pCO2, pH, Hct, BE, Hb, SvO2, SaO2, and rO2, recorded during CPB. Partial (ie, between surgical phases) and total CPB mean values of each parameter were correlated with the GFAP plasma concentrations. Correlations are reported in Table 3; GFAP increase showed a tendency to correlate with NRTI (P ¼ 0.056, Spearman’s r ¼ 0.695), whereas it did not correlate with age, weight, CPB, aortic cross-clamp, DHCA, regional cerebral perfusion, minimum temperature, hypothermia duration, and rate of rewarming. Discussion Early neurologic complications have emerged as a key issue in the long-term quality of life in children undergoing CHD surgery during infancy. Several factors have been suspected to contribute to increasing neurologic deficit, but no single factor has ever been recognized as the predominant one. Preoperative brain insults may occur since fetal life, especially in CHD with a diminished cerebral flow and brain dysgenesis (ie, univentricular physiology), continuing in the early postnatal life with cyanosis, cerebral circulation dysregulation, and hypoxic-ischemic insults.17 Moreover, the same techniques responsible for enhanced survival in CHD surgery (low-flow CPB and DHCA) may also be responsible for the brain injury associated with surgical techniques that require periods of decreased or interrupted cerebral flow.18 Finally, blood pH variations during CPB and DHCA are responsible for impaired cerebral vasoregulation resulting in increased vascular resistance combined with a diminished blood flow, which can cause brain damage.8,9 Several strategies, such as regional cerebral perfusion either antegrade or retrograde, were developed with the aim to overcome or minimize CHD surgery-related injuries. However, neurologic outcomes of children undergoing CHD surgery seem unchanged.10,12,14 Cerebral injury biomarkers are a promising tool to identify brain damage during or early after surgery. Neuronal enolase,13 S100b,19 HSP-27 and 70,20 pNF-H,21 and UCHL-122 have been investigated as markers of cerebral damage. Despite some of these biomarkers having interesting diagnostic properties, GFAP seems to fully

358

Canadian Journal of Cardiology Volume 32 2016

Table 1. Patients’ clinical and surgical characteristics Patient’s characteristics

Group 1

Clinical Median age (mo) Median weight (kg) Gender (m/f) Preterm birth (n) Oxygen saturation presurgery (%) Haemoglobin Base excess Cerebral saturation presurgery pO2 presurgery (%) pCO2 presurgery pH presurgery Lactate presurgery Post-op survival rate (%) Surgery Duration of surgery (min) CPB time (min) CPB flow (mL/kg/min) Aortic cross-clamp time (min) DHCA (n) DHCA time (min) Regional cerebral perfusion Duration (min) Flow (mL/kg/min) Temperature ( C) Minimum temperature ( C) Minimum temperature interval (min) Rewarming rate ( C/min) NRTI (min) Cerebral saturation < 45% during CPB (% of CPB time) Conversion to ECMO (n) Cerebral saturation difference during CPB

0.45 (0.23-0.92) 3.3 (2.9-3.4) 4/4 1/8 89 (86-90) 11.7 (9.8-13.3) 4.5 (3.1-7.9) 58.6 (33.4-65.5) 111 42 7.45 1.55 87.5%

(60-156) (29-46) (7.34-7.48) (0.76-2.98) (7/8)

315 180 123 37.5 37.5% 29 87.5% 48.5 44.9 24.0 23.0 20.0 0.33 51.5 40.0 37.5% 33

(289-382) (165-217) (111-128) (23.5-46.2) (3/8) (18-32) (7/8) (35.8-55.8) (31.7-54.0) (18.5-25.2) (19.2-24.7) (12.5-30.0) (0.18-0.73) (36.5-60.3) (10.0-69.0) (3/8) (19-43)

Group 2

P

48.7 (10.5-49.3) 13.7 (9.7-14.6) 2/3 1/5 86 (82-97) 14.0 (12.9-17.2) 4.0 (6.4-3.4) 65.2 (59.5-78.3)

0.003 0.003

73 37 7.36 0.75 100%

0.941 0.034 0.107 0.143

(50-89) (32-48) (7.26-7.40) (0.67-1.07) (5/5)

0.188 0.661 0.040 0.188

230 (220-252) 99 (75-119) 97 (134-144) n.a. 0/5 n.a. n.a. n.a. n.a. n.a. 33.0 (31.5-34.2) 40.0 (25.0-50.0) 0.15 (0.10-0.25) n.a. 7.1 (0-22.5) 0% (0/5) 24 (6-26)

0.027 0.013 0.732

0.003 0.136 0.057 0.077 0.092

CPB, cardiopulmonary bypass; DHCA, deep hypothermic cardiac arrest; ECMO, extracorporeal membrane oxygenation; NRTI, neurologic risk time interval.

fit the properties required by a diagnostic brain injury marker.11,15 In fact, GFAP is a fundamental constituent of the cytoskeleton of astrocytes, is readily released proportionally to the degree of injury into the bloodstream after various brain insults like stroke,23 traumatic brain injury,24 and cardiac arrest.25 It is not constitutively expressed, and thus it is probably released only after cell death.26 Among CHDs, univentricular hearts carry a burden of the highest risk of neurologic, cognitive, and psychological impairment27 that can be related to preoperative conditions, prolonged cyanosis, or multiple surgical palliation stages.28 In this initial report, we have analyzed the variation of GFAP as glial damage marker during CPB phases in children with univentricular physiology CHD who underwent

first-stage surgical palliation in neonatal age, or the Fontan procedure in infancy. We selected patients who were on stable hemodynamic conditions and similar GFAP preliminary status (GFAP nondetectable at the beginning of the study), so as

Table 2. Glial fibrillary acidic protein status in study groups during surgery Surgical stage Presurgery CPB start End of hypothermia End of rewarming End of CPB Postprotamine Maximum AUC (units)

Aortic clamp

1.36 1.87 1.24 0.04 1.67 240

No aortic clamp

n.d. n.d. (0.50-3.34) (0.89-4.36) (0.83-4.04) (0.02-0.52) (0.98-5.28) (152-570)

AUC, area under the curve; CPB, cardiopulmonary bypass.

n.d. n.d. n.d. n.d n.d n.d. n.d. n.d.

Figure 1. Group 1 median, interquartile range, and full range of plasma GFAP at different cardiopulmonary bypass stages. The outlier sample measured at the end of hypothermia is commented in the text. CPB, cardiopulmonary bypass; GFAP, glial fibrillary acidic protein.

Vedovelli et al. GFAP in Univentricular CHD

359

Table 3. Correlations of glial fibrillary acidic protein at specific surgery stages with clinical parameters in the aortic cross-clamp group Spearman’s P

r

0.028

0.762

% rO2 < 45% (total CPB mean) End of hypothermia 0.021 End of rewarming 0.052 End of CPB 0.028 AUC 0.021 Max 0.007

0.786 0.750 0.762 0.786 0.857

Neurologic risk time interval (NRTI)

AUC

0.056

0.695

pH (total CPB mean)

End of rewarming End of CPB

0.023 0.037

0.821 0.738

Parameter Cerebral saturation (rO2) difference during CPB

Surgical stage End of CPB

% rO2 < 45% ¼ percentage of CPB time spent under 45% of rO2 (cerebral regional oxygen saturation). AUC, area under the curve; CPB, cardiopulmonary bypass.

to investigate the role of the CPB stages in brain injury. We found that increased GFAP concentrations occur at the end of hypothermia, at the end of rewarming, and at the end of CPB only in children who have undergone aortic cross-clamp and had an NRTI during surgery (group 1). No GFAP variation occurred in children who underwent the Fontan procedure that did not require aortic cross-clamp or an NRTI (group 2). At this stage, we cannot establish if a GFAP increase is caused by the abnormal cerebral blood flow during NRTI or by other concomitant factors. In addition, we noticed that the GFAP plasma level started to increase at the end of hypothermia, which also corresponds to a decrease in cerebral blood flow. Then a peak was registered at the end of the rewarming phase, and finally it decreased to baseline values after CPB. These findings confirm previous studies that reported rewarming as the most susceptible phase of CPB for the GFAP increase during CHD surgery.15 The overall highest GFAP concentration was found at the end of hypothermia in one 7-month-old patient, a former preterm (26þ5 gestational age), who underwent Damus-KayeStansel palliation after a neonatal banding of pulmonary artery branches. The patient was maintained on deep hypothermia for 50 minutes (2.5 times longer than the mean duration in other patients of the same group), and on regional cerebral perfusion

for 54 minutes (in line with the mean of the group). She also underwent the fastest rate of rewarming (1.2 vs 0.36 C/min). As a single case, we could not ascribe to one single factor the cause of this abrupt increase. The rank-based, nonparametric statistical analysis permitted the inclusion of this patient into the analysis without the risk of a leverage effect of the outliers. We speculated that the GFAP level could reflect a direct insult against white matter oligodendrocyte progenitors similar to the onset of periventricular white matter injury of preterms, which could be predicted by GFAP blood levels measured soon after birth.29 In this way, less mature brain cells (first-stage palliation) may be more susceptible to an ischemic-hypoxic insult that is the proposed mechanism of white matter injury associated with CPB.30 We correlated GFAP concentrations of group 1 patients (undergoing aortic cross-clamp and NRTI, 8 of 13) at each surgical stage with their respective clinical and surgical data. GFAP strongly correlated with the cerebral desaturation expressed as the percentage of time spent on rO2 < 45% during CPB, regardless of the temperature (range, 18 C25.5 C) during the hypothermia phase. In particular, GFAP measured at the end of hypothermia and at the end of CPB showed the most significant correlation with the CPB rO2 < 45%. In addition, percentage of time of rO2 < 45% correlated with the GFAP peak and the GFAP AUC (Table 3). We also found that GFAP correlated with the absolute rO2 difference during CPB. Finally, GFAP AUC tended to be correlated (P ¼ 0.056 and Spearman’s r ¼ 0.695) with the NRTI (Fig. 2). These findings suggest that during NRTI, neurologic-risky manoeuvres or conditions could take place, independently from the applied technique. Mean CPB pH directly correlates with GFAP at the end of rewarming and at the end of CPB. Because alpha-stat pH management (ie, blood alkalosis was not corrected during hypothermia) was used for all patients, we speculate that alkalosis could have a role in the GFAP increase. Moreover, GFAP correlates with the increased pH but not with the hypothermia lowest temperature; thus we are more confident to attribute to alkalosis a role in a GFAP increase instead of a secondary effect of the temperature. Lastly, no correlations were found between GFAP and age, weight, length of surgical procedure, CPB duration, crossclamp duration, minimum temperature, hypothermia

Figure 2. Glial fibrillary acidic protein (GFAP) area under the curve (AUC) correlation (simple linear regression) with the neurologic risk time interval (NRTI). The time spent under neurologic-risky conditions (selective cerebral perfusion and/or deep hypothermic circulatory arrest). (A) All patients included. P ¼ 0.18, r2 ¼ 0.28. (B) Preterm outlier excluded. P ¼ 0.01, r2 ¼ 0.72. The outlier at around 1000 units is described in the discussion.

360

duration, DHCA duration, and rate of rewarming, suggesting that, in the range used in our cohort, these parameters are not associated with the increase of GFAP or, conversely, patients number was too scarce to notice significant differences. No difference in GFAP levels was found in patients (3 of 8) who had a direct conversion from CPB to extracorporeal membrane oxygenation. None of the patients had major neurologic impairment (ie, seizures and/or motor deficit) at discharge. A major limitation of this study is the small number of patients who did not permit a more detailed statistical analysis. In addition, because the age and weight of the 2 groups were significantly different, we cannot exclude a role of age and brain maturation. Older age and more mature brain could be less susceptible to the cardiac surgical injury, although the surgery was accomplished on CPB without cross-clamp. Because different neuroprotective strategies do not seem to correlate with neurodevelopmental impairment,31 we decided to include in the study patients who underwent similar but different surgeries involving CPBs with different characteristics. Thus, we focused more on the CPB phases than on the neuroprotective strategies. Moreover, the study was single centred and comprised a small number of patients who were also not compared with matched children with CHD with biventricular physiology. Finally, our study did not address the vascular autoregulation status during surgery that seems to be a useful NIRS-derived parameter that correlates with the GFAP increase.32 Conclusions Overall data in our study in infants and children suggest that patients with CHD and univentricular physiology who undergo surgical procedures with aortic cross-clamp and NRTI are related to a significant increase in the GFAP plasma level, which is a reliable brain (white matter) injury biomarker. On the contrary, late palliative surgeries with CPB and no aortic cross-clamp, or regional cerebral perfusion nor NRTI, do not affect the plasma concentration of GFAP at any time of the surgery. In our experience, GFAP increases along with prolonged cerebral hypoxia and, in general, with marked changes in cerebral oxygen saturation. Alkalosis also seems to be a good secondary indicator of the GFAP increase. A comparison of the effect of alpha-stat vs pH-stat management on GFAP concentration could be interesting to address the role of pH in brain injury during surgery. A larger population and clinical evaluation of these patients at follow-up may possibly confirm the reliability of GFAP as a brain injury marker and its neurodevelopmental clinical implications. Acknowledgements The authors would like to thank Eric Hill for the language support. Funding Sources This study was supported by the “Grant Program for Young Investigators on Pediatric Research” of the Fondazione Cassa di Risparmio di Padova e Rovigo 2013, granted to L.V., and by the “Ricerca Corrente 2013, 2014” Grants from the Bambino Gesù Children’s Hospital, granted to P.E.C. The funding sources had no involvement in any part of the study.

Canadian Journal of Cardiology Volume 32 2016

Disclosures The authors have no conflicts of interest to disclose. References 1. Hoffman JI, Kaplan S. The incidence of congenital heart disease. J Am Coll Cardiol 2002;39:1890-900. 2. Reller MD, Strickland MJ, Riehle-Colarusso T, Mahle WT, Correa A. Prevalence of congenital heart defects in metropolitan Atlanta, 19982005. J Pediatr 2008;153:807-13. 3. Biglino G, Giardini A, Hsia TY, et al. Modeling single ventricle physiology: review of engineering tools to study first stage palliation of hypoplastic left heart syndrome. Front Pediatr 2013;1:31. 4. Dancea A, Justino H, Martucci G. Catheter intervention for congenital heart disease at risk of circulatory failure. Can J Cardiol 2013;29:786-95. 5. Agarwal HS, Wolfram KB, Saville BR, Donahue BS, Bichell DP. Postoperative complications and association with outcomes in pediatric cardiac surgery. J Thorac Cardiovasc Surg 2014;148:609-16. e601. 6. Mondesert B, Marcotte F, Mongeon FP, et al. Fontan circulation: success or failure? Can J Cardiol 2013;29:811-20. 7. Majnemer A, Limperopoulos C, Shevell M, et al. Developmental and functional outcomes at school entry in children with congenital heart defects. J Pediatr 2008;153:55-60. 8. du Plessis AJ. Mechanisms of brain injury during infant cardiac surgery. Semin Pediatr Neurol 1999;6:32-47. 9. Albers EL, Bichell DP, McLaughlin B. New approaches to neuroprotection in infant heart surgery. Pediatr Res 2010;68:1-9. 10. Bellinger DC, Jonas RA, Rappaport LA, et al. Developmental and neurologic status of children after heart surgery with hypothermic circulatory arrest or low-flow cardiopulmonary bypass. N Engl J Med 1995;332:549-55. 11. Eng LF, Ghirnikar RS, Lee YL. Glial fibrillary acidic protein: GFAPthirty-one years (1969-2000). Neurochem Res 2000;25:1439-51. 12. Hirsch JC, Jacobs ML, Andropoulos D, et al. Protecting the infant brain during cardiac surgery: a systematic review. Ann Thorac Surg 2012;94: 1365-73 [discussion: 1373]. 13. Schmitt B, Bauersfeld U, Schmid ER, et al. Serum and CSF levels of neuron-specific enolase (NSE) in cardiac surgery with cardiopulmonary bypass: a marker of brain injury? Brain Dev 1998;20:536-9. 14. Wernovsky G, Shillingford AJ, Gaynor JW. Central nervous system outcomes in children with complex congenital heart disease. Curr Opin Cardiol 2005;20:94-9. 15. Brunetti MA, Jennings JM, Easley RB, et al. Glial fibrillary acidic protein in children with congenital heart disease undergoing cardiopulmonary bypass. Cardiol Young 2014;24:623-31. 16. Donmez A, Yurdakok O. Cardiopulmonary bypass in infants. J Cardiothorac Vasc Anesth 2014;28:778-88. 17. Gaynor JW. The encephalopathy of congenital heart disease. J Thorac Cardiovasc Surg 2014;148:1790-1. 18. Lynch JM, Buckley EM, Schwab PJ, et al. Time to surgery and preoperative cerebral hemodynamics predict postoperative white matter injury in neonates with hypoplastic left heart syndrome. J Thorac Cardiovasc Surg 2014;148:2181-8.

Vedovelli et al. GFAP in Univentricular CHD

361

19. Erb MA, Heinemann MK, Wendel HP, et al. S-100 after correction of congenital heart defects in neonates: is it a reliable marker for cerebral damage? Ann Thorac Surg 2000;69:1515-9.

26. Andropoulos DB, Brady KM, Easley RB, Fraser CD Jr. Neuroprotection in pediatric cardiac surgery: what is on the horizon? Prog Pediatr Cardiol 2010;29:113-22.

20. Franklin TB, Krueger-Naug AM, Clarke DB, Arrigo AP, Currie RW. The role of heat shock proteins Hsp70 and Hsp27 in cellular protection of the central nervous system. Int J Hyperthermia 2005;21:379-92.

27. Kaulitz R, Hofbeck M. Current treatment and prognosis in children with functionally univentricular hearts. Arch Dis Child 2005;90:757-62.

21. Shaw G, Yang C, Ellis R, et al. Hyperphosphorylated neurofilament NFH is a serum biomarker of axonal injury. Biochem Biophys Res Commun 2005;336:1268-77.

28. Beca J, Gunn JK, Coleman L, et al. New white matter brain injury after infant heart surgery is associated with diagnostic group and the use of circulatory arrest. Circulation 2013;127:971-9.

22. Massaro AN, Jeromin A, Kadom N, et al. Serum biomarkers of MRI brain injury in neonatal hypoxic ischemic encephalopathy treated with whole-body hypothermia: a pilot study. Pediatr Crit Care Med 2013;14: 310-7.

29. Stewart A, Tekes A, Huisman TA, et al. Glial fibrillary acidic protein as a biomarker for periventricular white matter injury. Am J Obstet Gynecol 2013;209:27. e21-27.

23. Dvorak F, Haberer I, Sitzer M, Foerch C. Characterisation of the diagnostic window of serum glial fibrillary acidic protein for the differentiation of intracerebral haemorrhage and ischaemic stroke. Cerebrovas Dis 2009;27:37-41.

30. Agematsu K, Korotcova L, Scafidi J, et al. Effects of preoperative hypoxia on white matter injury associated with cardiopulmonary bypass in a rodent hypoxic and brain slice model. Pediatr Res 2014;75:618-25.

24. Nylen K, Ost M, Csajbok LZ, et al. Increased serum-GFAP in patients with severe traumatic brain injury is related to outcome. J Neurol Sci 2006;240:85-91.

31. Wypij D, Newburger JW, Rappaport LA, et al. The effect of duration of deep hypothermic circulatory arrest in infant heart surgery on late neurodevelopment: the Boston Circulatory Arrest Trial. J Thorac Cardiovasc Surg 2003;126:1397-403.

25. Kaneko T, Kasaoka S, Miyauchi T, et al. Serum glial fibrillary acidic protein as a predictive biomarker of neurological outcome after cardiac arrest. Resuscitation 2009;80:790-4.

32. Rhee CJ, Kibler KK, Brady KM, et al. Detection of neurologic injury using vascular reactivity monitoring and glial fibrillary acidic protein. Pediatrics 2013;131:e950-4.