PEDIATRIC ANESTHESIA SOCIETY FOR PEDIATRIC ANESTHESIASECTION EDITORWILLIAM J. GREELEY

A Glial-Derived Protein, S100B, in Neonates and Infants withCongenital Heart Disease: Evidence for PreexistingNeurologic InjuryPaula M. Bokesch, MD*, Elumalai Appachi, MD, Marco Cavaglia, MD*, Emad Mossad, MD*,and Roger B.B. Mee, MB, ChB, FRACSDepartments of *Cardiothoracic Anesthesia, Pediatric Critical Care, and the Center for Congenital Heart Disease andSurgery, The Cleveland Clinic Foundation, Ohio

The glial-derived protein S100B is a serum marker of cere-bral ischemia and correlates with negative neurologicaloutcome after cardiopulmonary bypass (CPB) in adults.We sought to characterize the S100B release pattern beforeand after CPB in neonates and infants with congenitalheart disease and correlate it with surgical mortality. Se-rum was collected before surgery and at 24 postoperativeh from 109 neonates and infants with congenital heart dis-ease. All patients had presurgical transthoracic echocar-diograms and CPB with or without hypothermic circula-toryarrest.S100Bconcentrationsweredeterminedusingatwo-site immunoluminometric assay (Sangtec 100).Thirty-day surgical mortality was observed. All neonateshad significantly increased S100B concentrations before

surgery that decreased by 24 postoperative h. Preopera-tive S100B concentrations in 32 neonates with hypoplasticleft heart syndrome correlated inversely with the forwardflow and size of the ascending aorta and postoperativemortality (r2 !"0.63; P!0.03). Among infants, increasedpulmonary blood flow was associated with higher S100Blevels before surgery than cyanosis. There was no correla-tion with postoperative S100B and time on CPB, hypo-thermic circulatory arrest, or 30-day surgical mortality. Inconclusion, preoperative S100B concentrations correlateinversely with the size of the ascending aorta in hypoplas-tic left heart syndrome and may serve as a marker for pre-existing brain injury and mortality.

(Anesth Analg 2002;95:88992)

Improved techniques in cardiopulmonary bypass(CPB) combined with deep hypothermia and hypo-thermic circulatory arrest have enabled correctionof complex congenital heart defects in the neonataland early infancy periods. Although advances in tech-nological, surgical, and anesthetic techniques have im-proved morbidity and mortality for these critically illbabies, developmental and neurological abnormalitiesin follow-up studies may still occur in neonates, in-fants, and children (13).

Neurological dysfunction after heart surgery is usu-ally attributed to intraoperative ischemic events occur-ring during CPB, hypothermic circulatory arrest(HCA), and after reperfusion and rewarming (1,4).Another important factor that may contribute to the

neurologic morbidity after open-heart surgery is apreexisting brain abnormality secondary to the de-fective cardiac anatomy and physiology in utero andat birth. Brain and developmental abnormalities arevery difficult to detect in a newborn baby. Neuro-psychological examinations are not useful or feasi-ble in a critically ill neonate. Imaging techniques orcranial ultrasounds do not provide functional data.A biochemical marker of cerebral injury is highlydesirable both before CPB to detect and quantifypreexisting brain injury as well as to determine theextent of cerebral injury in association with CPB andHCA.

The glial-derived protein S100B is a serum markerof cerebral ischemia. Patients suffering from stroke,subarachnoid hemorrhage, and head trauma releaseS100B into the cerebrospinal fluid and blood (5,6).In adults, serum concentrations of S100B afterCPB correlate with detrimental neurological out-come (7). We sought to characterize the S100B re-lease pattern before and after CPB in neonates andinfants and determine its relationship to patientmortality.

Presented by Dr. Julie Tome, Department of Cardiothoracic An-esthesia, at the Annual Meeting of the American Society of Anes-thesiologists, New Orleans, LA, October 2001.

Accepted for publication June 3, 2002.Address correspondence and reprint requests to Paula M.

Bokesch, MD, Department of Cardiothoracic Anesthesia, G30 Cleve-land Clinic Foundation, 9500 Euclid Ave., Cleveland, OH 44195.Address e-mail to bokescp@ccf.org.

DOI: 10.1213/01.ANE.0000026490.66200.43

2002 by the International Anesthesia Research Society0003-2999/02 Anesth Analg 2002;95:88992 889

MethodsAfter IRB approval and parental informed consent,109 term neonates or infants undergoing electivecardiac surgery using CPB were sequentially en-rolled in this study. Patients with Down syndromeand neonates who suffered a cardiac arrest, hypo-tension, or severe acidosis, with a pH value #7.2,within 24 h before surgery were excluded. Patientswere grouped by type of repair: Stage I repair ofhypoplastic left heart syndrome (HLHS, Group 1),transposition of the great arteries (Group 2), tetral-ogy of Fallot (Group 3), bi-directional cavopulmo-nary anastomosis (Group 4), or ventricular septaldefects (VSD, Group 5). All neonates had preoper-ative renal and head ultrasounds before surgery asper standard procedure at Cleveland Clinic Foun-dation. Only those with normal ultrasounds wereincluded in this study. All patients had preoperativeechocardiograms that determined chamber andvalve size and antegrade or retrograde flow in theaorta. Demographic, intra-, and postoperative dataon each patient were collected from patient records.Surgical mortality for 30 days after surgery wasdocumented.

After the induction of anesthesia, all patients had4.55.0F double-lumen catheters placed in the internaljugular vein as per standard procedure for monitor-ing, drug infusion, and blood sample collection. Inmost patients, a right atrial catheter inserted by thesurgeon replaced this catheter at the end of surgery.These catheters were used for venous sampling toassay S100B concentrations. One milliliter of venousblood was obtained from all patients before the inci-sion and 24 h after arrival in the pediatric intensivecare unit (ICU). The samples were immediately cen-trifuged and the serum stored at "80C until analysis.

Serum S100B concentrations were determined usinga monoclonal two-site immunoluminometric assay ac-cording to the manufacturers directions (LIA-MATSangtec 100, AB Sangtec Medical, Bromma, Swe-den). The precision was #10.0% coefficient of varia-tion. The lower level of detection was 0.1 !g/L, andthe upper level was 20.0 !g/L.

If not otherwise stated, all data are presented asmean $ sem. Data were analyzed using SigmaStat(Jandel Scientific, San Rafael, CA). For comparisonsof means, continuous variables were analyzed withStudents t-tests if the sample was large or had anormal distribution. Otherwise, a nonparametrictest was used (Mann-Whitney). Regression analysiswas performed using the least square method withcase-wise deletion of missing data. Correlation co-efficients were determined using the Pearson Prod-uct Moment Correlation. A P value #0.05 was con-sidered significant.

ResultsThe pre- and postoperative serum concentrations ofS100B are presented in Table 1. Group 1 patients withHLHS had the largest mean S100B concentrationsbefore and 24 h after surgery compared with allother groups (P # 0.01). Preoperative S100B concen-trations correlated significantly (P ! 0.003) and in-versely (r2 ! "0.64) with the size of the ascendingaorta and the presence of forward flow through theaortic valve documented by transthoracic echocar-diography (Fig. 1). All patients who died (n ! 5) hadpreoperative serum concentrations "3.6 !g/L andsmall ascending aorta diameters without forwardflow. None of these patients were acidotic at thetime of sampling or within 24 h of surgery. If thesepatients were excluded from data analysis, the meanpreoperative S100B concentration was 1.75 $ 1.7!g/L, which was similar to Group 2 patients.

As with Group 1, Group 2 patients with transpo-sition of the great arteries also had increased pre-operative S100B, as compared with Groups 3 and 4,that decreased 24 h after surgery (Table 1). Withinthis group, patients with VSD (n ! 17) did not havepreoperative concentrations of S100B significantlydifferent from patients with intact ventricular sep-tum (1.95 $ 1.7 !g/L versus 2.2 $ 1.4 !g/L, respec-tively; P ! 0.3). All infants (Groups 3 and 4) hadsignificantly smaller preoperative concentrations ofS100B than patients in Groups 1 and 2 (P # 0.01). How-ever, among infants, cyanotic infants (Groups 3 and 4)had significantly smaller preoperative S100B concentra-tions than noncyanotic infants with VSD and increasedpulmonary circulation (Group 5; P # 0.05).

Although S100B concentrations were larger at24 h than before surgery in Group 4, all of theseinfants were extubated and discharged from theICU on the first postoperative day. As expected, allof these patients had significantly (P # 0.01) in-creased pressure in the superior vena cava on ar-rival in the ICU (12.2 $ 2 mm Hg versus 5.6 $ 1 mmHg before CPB). There was no correlation betweenthe 24-h postoperative S100B concentrations and theduration of CPB or HCA, aortic cross-clamp time, orsurvival in any of the groups.

DiscussionS100B has been established as a useful serum markerof cerebral injury caused by minor and major headtrauma, global anoxia, focal ischemia, metastatic ma-lignant melanoma, and cardiac surgery (57). All ofthese conditions involve disruption of the blood-brainbarrier. Johnsson et al. (8) first reported the relation-ship of increased S100B and cerebral complicationsafter cardiac surgery in adults. These authors andothers subsequently determined that shed mediastinal

890 PEDIATRIC ANESTHESIA BOKESCH ET AL. ANESTH ANALGA GLIAL-DERIVED PROTEIN, NEUROLOGIC INJURY, AND HEART DISEASE 2002;95:88992

blood collected during surgery by cardiotomy suctionalso contained high levels of S100B as well as chesttube blood that was used for autotransfusion aftersurgery (9). Therefore, earlier reports that correlatedserum S100B levels immediately after pediatric car-diac operations with duration of CPB and HCA mayhave been contaminated by extra-cerebral sources ofS100B (10,11).

Our study did not find any relationship betweenCPB duration, cross-clamp time, or use of HCA andserum concentrations of S100B 24 hours after surgery.Also, serum concentrations of S100B at 24 hours aftersurgery did not correlate with 30-day surgical mortal-ity. Unlike adult studies, neonates and infants hadsmaller concentrations of S100B at 24 hours after sur-gery than before surgery. However, this finding mayreflect dilution of the protein in serum from postop-erative blood, colloid, and crystalloid infusions insmall babies.

Although neonates in Groups 1 and 2 had largerS100B levels than previously reported in adults, thesevalues were not necessarily abnormal. Age-related se-rum S100B concentrations before surgery have beendescribed in children with congenital heart disease

with the highest values in neonates (12,13). Erb et al.(14) reported serum concentrations of 1.97 !g/L in12 healthy neonates on the day of birth. Maschmann etal. (15) reported serum S100B concentrations in 66term newborns to be 0.663.33 !g/L in the first weekof life. Newborns with signs of hypoxic-ischemic en-cephalopathy after perinatal acidosis had increasedserum concentrations of S100B %4.0 !g/L in the firstweek of life. However, compared with serum levels ofadult patients (usually #0.2 !g/L), even nonacidoticnewborns have significantly larger S100B concentra-tions. The reason for the higher S100B serum levels inhealthy newborns is unclear. Some possibilities areimmaturity of the blood-brain barrier with subsequentleakage of S100B from the brain into the blood, extra-cerebral sources of S100B such as brown fat, or de-creased elimination by immature kidneys. In adults,S100B is eliminated, unchanged by the kidneys with ahalf-life of 25.3 $ 5.1 minutes after CPB, and is notaffected by a moderate decrease in glomerular filtra-tion rate (9).

The most intriguing observation in our study wasthe relationship between preoperative S100B concen-trations and the size of the ascending aorta normalizedto body surface area in neonates undergoing the Nor-wood operation (Stage I repair of HLHS). Smallerascending aorta correlated with larger concentrationsof S100B (Fig. 1; P ! 0.003). Patients with Shonescomplex (mitral stenosis and aortic stenosis withoutatresia) and other variants of HLHS, who had forwardflow through their aortic valve (confirmed by preop-erative echocardiography), had smaller preoperativeconcentrations of S100B. These data suggest that neo-nates with little or no forward flow in the ascendingaorta have brain injury before surgery. None of theseneonates were acidotic or hypotensive at this time.

Group 5 infants had a similar finding as Group 1neonates, namely decreased forward flow in the as-cending aorta. Patients with VSDs have excessive pul-monary circulation through the VSD, thereby decreas-ing forward flow in the ascending aorta. Age-matchedcyanotic patients with normal or increased flow pat-terns in the ascending aorta (Groups 3 and 4) hadsmaller preoperative S100B concentrations.

Figure 1. Preoperative [S100B] versus ascending aorta diameternormalized to body surface area. Dark squares (5) are neonates whodied after Stage I repair of HLHS. r2 !"0.64; P ! 0.003; n ! 32 StageI repair of HLHS.

Table 1. Pre- and Postoperative Serum S100B Concentrations

Group

S100B !g/L

Preop 24 h Postop

1 Stage I HLHS (32) 8.6 $ 1.2 days 2.5 $ 0.4* 2.0 $ 0.42 TGA Switch (28) 10.4 $ 3.8 days 2.1 $ 0.4* 1.5 $ 0.33 TOF Infants (10) 6.0 $ .8 months 0.5 $ 0.3 0.5 $ 0.44 BDCP Infants (21) 5.1 $ 1.4 months 0.4 $ 0.1 0.9 $ 0.35 VSD Infants (18) 5.8 $ 1.2 months 1.1 $ 0.2 1.0 $ 0.2

HLHS ! hypoplastic left heart syndrome; TGA ! transposition of the great arteries; TOF ! tetralogy of Fallot; BDCP ! bidirectional cavo-pulmonaryanastomosis; VSD ! ventricular septal defect; preop ! preoperative; postop ! postoperative.

* P # 0.01 Groups 1 and 2 versus Groups 34 preop; P # 0.05 postop versus preop Group 4; P # 0.05 Group 5 versus Groups 3 and 4.

ANESTH ANALG PEDIATRIC ANESTHESIA BOKESCH ET AL. 8912002;95:88992 A GLIAL-DERIVED PROTEIN, NEUROLOGIC INJURY, AND HEART DISEASE

Alessandro Alfieri

Alessandro Alfieri

Alessandro Alfieri

Group 4 infants were all previous Stage I repair ofHLHS and its variants. At 5.8 $ 2 months of age, theseinfants had significantly smaller concentrations ofS100B before surgery than they did as neonates. Thesignificant increase of S100B after surgery in thisgroup may reflect the increased pressure in the supe-rior vena cava, although more patients are required todetermine if this correlation is consistent.

In summary, preoperative serum concentrations ofS100B are increased in neonates and infants with con-genital heart disease. Preoperative serum S100B con-centrations correlate inversely with the forward flowand size of the ascending aorta in HLHS and mayserve as an indicator of preoperative brain injury anda predictor of survival.

References1. Bellinger DC, Wypij D, Kuban KC, et al. Developmental and

neurological status of children at 4 years of age after heartsurgery with hypothermic circulatory arrest or low-flow cardio-pulmonary bypass. Circulation 1999;100:52632.

2. Hovels-Gurich HH, Seghaye MC, Dabritz S, et al. Cognitive andmotor development in preschool and school-aged children afterneonatal arterial switch operation. J Thorac Cardiovasc Surg1997;114:57885.

3. Miller G, Eggli KD, Contant C, et al. Postoperative neurologiccomplications after open heart surgery on young infants. ArchPediatr Adolesc Med 1995;149:7648.

4. Taylor KM. Central nervous system effects of cardiopulmonarybypass. Ann Thorac Surg 1998;66:S204.

5. Raabe A, Grolms C, Sorge O, et al. Serum S-100B protein insevere head injury. Neurosurgery 1999;45:47783.

6. Wunderlich MT, Ebert AD, Kratz T, et al. Early neurobehavioraloutcome after stroke is related to release of neurobiochemicalmarkers of brain damage. Stroke 1999;30:11905.

7. Georgiadis D, Berger A, Kowatschev E, et al. Predictive value ofS100B and neuron-specific enolase serum levels for adverseneurologic outcome after cardiac surgery. J Thorac CardiovascSurg 2000;119:13847.

8. Johnsson P, Lundqvist C, Lindgren A, et al. Cerebral complica-tions after cardiac surgery assessed by S-100 and NSE levels inblood. J Cardiothorac Vasc Anesth 1995;9:6949.

9. Jonsson H, Johnsson P, Hoglund P, et al. Elimination of S100Band renal function after cardiac surgery. J Cardiothorac VascAnesth 2000;14:698701.

10. Taggart DP, Mazel JW, Bhattacharya K, et al. Comparison ofserum S-100B levels during CABG and intracardiac operations.Ann Thorac Surg 1997;63:4926.

11. Abdul-Khaliq H, Blasig IE, Baur MO, et al. Release of the cere-bral protein S-100B into blood after reperfusion during cardiacoperations in infants: is there a relation to oxygen radical-induced lipid peroxidation? J Thorac Cardiovasc Surg 1999;117:10278.

12. Lindberg L, Olsson AK, Anderson K, Joegi P. Serum S-100protein levels after pediatric cardiac operations: a possible newmarker for postperfusion cerebral injury. J Thorac CardiovascSurg 1998;116:2815.

13. Jensen E, Sandstroem K, Andreasson S, et al. Increased levels ofS-100 protein after cardiac surgery with cardiopulmonary by-pass and general surgery in children. Paediatr Anaesth 2000;10:297302.

14. Erb MA, Heinemann MK, Wendel HP, et al. S100 after correc-tion of congenital heart defects in neonates: is it a reliablemarker for cerebral damage? Ann Thorac Surg 2000;69:15159.

15. Maschmann J, Erb MA, Ziemer G, Speer CP. Evaluation of S-100serum concentrations in healthy newborn and seven newbornswith perinatal acidosis. Acta Paediatr 2000;89:5535.

892 PEDIATRIC ANESTHESIA BOKESCH ET AL. ANESTH ANALGA GLIAL-DERIVED PROTEIN, NEUROLOGIC INJURY, AND HEART DISEASE 2002;95:88992