| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Technical Briefs |
1 Department of Pediatrics, Obstetrics and Reproductive Medicine, University of Siena, Siena, Italy; 2 Department of Cardiac Surgery S. Donato Milanese University Hospital, San Donato Milanese, Italy; 3 Department of Pediatrics and Neuroscience, G. Gaslini Children’s Hospital, Genoa, Italy; 4 Department of Maternal Fetal and Neonatal Health G. Garibaldi Hospital, Catania, Italy
aAddress correspondence to this author at: Department of Maternal Fetal and Neonatal Health, G. Garibaldi Hospital, Via Palermo 636, I-95100 Catania, Italy; fax 39-95-7595208, e-mail dgazzolo{at}hotmail.com
Abstract
Background: Ischemic-reperfusion injury of the brain is a major adverse event after cardiac surgery, especially when extracorporeal circuits are used. Because brain injury induces local overproduction of activin A, we measured plasma concentrations in children after open heart surgery with cardiopulmonary bypass (CPB) to investigate the potential of measuring activin A for early identification of infants at risk for brain damage.
Methods: We evaluated 45 infants (age <1 year) with congenital heart defects: 36 without overt neurologic injury, and 9 with neurologic injury on day 7 after the surgical procedure. Blood samples were taken before surgery, during surgery before CPB, at the end of CPB, at the end of surgery, and at 12 h after surgery. Neurologic development was assessed before surgery and on postoperative day 7.
Results: Activin A concentrations increased significantly during surgery (P <0.0001) to a maximum at the end of CPB. Infants who developed abnormal neurologic sequelae had concentrations significantly higher (P <0.0001, all comparisons) than patients with normal neurologic outcome at all evaluated times, but not before surgery. Activin A had a sensitivity of 100% (95% CI, 66%–100%) and a specificity of 100% (95% CI, 90%–100%) as a single marker for predicting neurologic abnormalities (area under the ROC curve, 1.0).
Conclusions: Activin A increases in children who experience poor neurologic outcomes after open heart surgery, and its assay may help in early identification of infants at risk for brain damage.
Open heart surgery with cardiopulmonary bypass (CPB) is associated with delayed motor development and neurologic abnormalities in children (1). Despite accurate intraoperative monitoring during CPB, brain damage may occur at a subclinical stage, with sedation effects hiding clinical symptoms and monitoring variables (2)(3). The brain produces several factors after different types of injuries (3), and production of activin A has been observed both in vitro and in vivo (4). Activin A, a glycoprotein produced in the central nervous system (5), enhances the survival of midbrain and hippocampal neurons (4)(5)(6)(7), decreases ischemic brain injury in infant rats (4)(8), and shields striatal and midbrain neurons against neurotoxic damage (4)(9). Activin A concentrations are highest in preterm newborns after perinatal hypoxia (10), in infants experiencing intraventricular hemorrhage (11), and in perinatal asphyxia (12). We investigated measurement of the low relative molecular mass form of activin A (Mr = 26 000) in blood as a method to monitor possible cerebral distress during CPB in children undergoing cardiac surgery for repair of congenital heart defects.
From January 2001 to June 2004, we conducted a case-control study of 45 infants (23 males and 22 females) who had no preexisting neurologic disorders or other comorbidities and were admitted to our referral centers for pediatric cardiac diseases [tetralogy of Fallot (n = 22), transposition of the great arteries (n = 9), aortic stenosis (n = 7), tricuspid atresia (n = 2), atrioventricular canal (n = 3), and multiple interventricular septal defect (Swiss cheese type; n = 2)]. The 36 infants with no overt postoperative neurologic injury represented the group with no brain injury, and the remaining patients (n = 9), who developed neurologic injury on follow-up day 7, represented the brain-damage group.
Heparinized samples of blood taken at 4 time points [before surgery (time 0), during surgery before CPB (time 1), at the end of CPB (time 2), at the end of surgery (time 3), and 12 h after surgery (time 4)] were assayed for activin A. The blood samples were centrifuged, and the plasma fractions were divided into several aliquots and stored at –20 °C until assay. We recorded laboratory and standard monitoring variables (peripheral and nasopharyngeal temperatures, pump flow rate, mean arterial and left/right atrial blood pressures, arterial blood pH, carbon dioxide and oxygen partial pressures, bicarbonate concentrations, base excess, glycemia, ion concentrations, oxygen saturation of tissue, and heart rate) at all perioperative time points. Informed consent was obtained before patient inclusion in the study, which was approved by the local human-investigation committee. Anesthetic and standardized CPB (
-stat regimen) techniques were performed according to previously published protocols (1).
Neurologic development was assessed preoperatively and on postoperative day 7 by the same examiner (all cases admitted to the study were followed at the Department of Cardiac Surgery, S. Donato Milanese University Hospital) with the Amiel-Tison Neurological Assessment at Term test, a consolidated neurologic test applied in the neonatal intensive care unit (NICU) and pediatric intensive care unit (13). An assessment of neurologic outcome was established for day 7, in accordance with discharge times for patients with no complications. Specifically, a single examiner at each center tested for resistance against passive movements, visual pursuit, reaching and grasping, and responses to visual and acoustic stimuli. Results were compared with those characteristic for infants of that age (in months) and scored as "normal" or "abnormal".
Activin A was measured by means of a specific 2-site enzyme immunoassay (Serotec) (11). The analytical detection limit of the activin A assay was <0.001 µg/L, and intraassay and interassay CVs were 2.5% and 3.0%, respectively, for concentrations between 0.5 µg/L and 4.5 µg/L. The activin A assay does not cross-react with inhibin A, inhibin B, follistatin, or activin B.
The Kolmogorov–Smirnov test showed experimental values to have a gaussian distribution, and data were expressed as the mean (SD). Statistical significance was assessed by one-way ANOVA for repeated measures (followed by the post hoc Tukey test for multiple comparisons) and the unpaired t-test when only 2 groups were compared. Pearson correlation coefficients were calculated to test the linear correlation of activin A concentration with CPB duration and with NICU stay (in days). We used ROC curve analysis (14) to estimate the probability of developing neurologic abnormalities according to the activin A concentration and compared this probability with the pretest probability, which was defined as the prevalence of brain damage in the entire group of infants (15). Statistical significance was assumed whenever P <0.05.
The values of clinical laboratory and standard monitoring variables recorded at the predetermined time points remained within the reference limits (P >0.05 for all variables; data not shown). Age, weight, male/female distribution, and CPB, clamping, and circulatory arrest durations did not differ between the 2 groups (Table 1
). Neurologic examinations on day 7 after the surgical procedure revealed neurologic abnormalities in 9 infants (brain-damage group: hypertonia/hypotonia syndrome, 6 patients; hemisyndrome, 3 patients) whose stays in the NICU were significantly longer (P <0.001) than in the group with no brain injury. Activin A was detectable in all samples (range, 0.20–3.2 µg/L). The lack of significant differences with respect to sex and gestational age at birth (P >0.05, all comparisons; data not shown) confirmed previous data (16). Concentrations increased significantly (P <0.0001) during surgery in both groups of children and reached the highest concentrations at the end of CPB in the group with no brain injury, and at the end of CPB and at the end of surgery in the children with poor neurologic outcomes (brain-damage group). Activin A concentrations at these time points were significantly higher (P <0.01 for all comparisons) than those of other time points (Fig. 1
). Concentrations at times 1–4 were significantly higher (P <0.0001) in children with brain damage than in those with normal neurologic outcomes on day 7 of follow-up (Fig. 1
). Finally, activin A concentration was significantly correlated with CPB duration [group with no brain injury, Pearson r = 0.974 (P <0.0001); brain-damage group, Pearson r = 0.913 (P <0.0001)] and NICU stay [Pearson r = 0.866 (P <0.0001); data not shown].
|
|
At cutoff values chosen by the ROC curve analysis (time 1, >0.94 µg/L; time 2, >1.4 µg/L; time 3, >0.85 µg/L), the activin A assay conducted in the perioperative periods demonstrated a sensitivity of 100% (95% CI, 66%–100%) and a specificity of 100% (95% CI, 90%–100%) for predicting neurologic abnormalities [area under the ROC curve, 1.0; 95% CI, 0.92%–1.0% for all time points (data not shown)]. Nine of 45 neonates developed neurologic abnormalities, for an overall prevalence of the diseases in the study population of 20% (95% CI, 8.4%–32%). This percentage was the predicted probability of a poor neonatal outcome before measuring the activin A concentration (pretest probability). At activin A concentrations greater than the thresholds defined by the ROC curve analysis, the probability of a poor outcome was as high as 100% (95% CI, 99%–100%), whereas if the activin A concentration remained unchanged, the probability was 0% (95% CI, 0%–1.9%; data not shown).
Our study provides evidence that the activin A concentration increases during CPB and correlates with CPB duration and NICU stay, suggesting a link between the increase in activin A and the degree of the cerebral distress. Open heart surgery with CPB is associated with brain injury (1). The activin A concentration in the brain increases in vitro after several cerebral injuries (4), and increased activin A secretion in vivo has been associated with cerebral hypoxia (10)(17), intraventricular hemorrhage (12), and hypoxic ischemic encephalopathy (10). Moreover, the evidence that activin A concentrations are higher in cerebrospinal fluid than in serum (18) and strongly correlated with markers of brain damage (19) indicates definite activin A production and secretion within the central nervous system. Our findings of increased activin A concentrations in the bloodstreams of infants during cardiac surgery with CPB suggest that the activin A concentration increases after cerebral distress. When peculiar hemodynamic adaptive mechanisms in the brain are present in children (1), CPB presumably is accompanied at least by a transient cerebral dysfunction and by the appearance of activin A in the blood, which is a specific response to operative stress and hemodynamic changes. The possibility that changes in the permeability of the blood-brain barrier mediated in part by temperature might play a role in increasing blood activin A concentrations must also be taken into account, however, because hypothermia determines transient perfusion defects of the cerebrum’s micro and macrocirculation, particularly within the thalamus and hippocampus (20).
Our study also has demonstrated the usefulness of activin A measurement for better identifying children at risk for a poor neurologic outcome when clinical and laboratory findings appear normal. Activin A concentrations were increased before CPB in children who later experienced brain damage, findings that suggest an increased susceptibility to surgical procedures in such cases. Monitoring activin A concentration at different times may have potential as a test for predicting poor neurologic outcome with 100% sensitivity and 100% specificity.
References
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
P <0.001 (vs times 0 and 3) and P <0.01 (vs time 4);