Between 1993 when aprotinin (Trasylol®; Bayer Pharmaceuticals Corporation, West Haven, CT) was first approved for use in the United States by the Food and Drug Administration and 2007 when it was suspended in the United States and other world markets, aprotinin was used intraoperatively in thousands of cardiac surgical procedures to reduce the risks of excessive bleeding and allogeneic blood transfusions associated with cardiopulmonary bypass (CPB). During that period, many studies in both adult and pediatric cardiac patients confirmed its efficacy in reducing blood loss and transfusion requirements.1–4 However, complications were not the primary focus of these investigations and were addressed only peripherally. In 2006, Mangano et al. undertook a prospective investigation to assess the safety profile of aprotinin. They found a dose-dependent doubling of the risk of renal failure requiring dialysis among adult patients who received aprotinin during primary or complex coronary artery surgery.5 These investigators also reported an increase in thrombotic/ischemic-related events, such as myocardial infarction and stroke, in aprotinin-treated patients undergoing primary surgery. In 2006, Karkouti et al. again found that adult cardiac patients treated with aprotinin developed a higher risk of postoperative renal dysfunction, defined as a 50% increase in serum creatinine (Cr) concentration during the first postoperative week.6 In 2007, an additional investigation by Mangano et al. found that adult cardiac patients receiving aprotinin had an increased 5-yr mortality after surgery.7 The above investigations along with the incriminating results of the Canadian BART study (Blood conservation using Antifibrinolytics: a Randomized Trial in a cardiac surgery population study) prompted Bayer Healthcare, under the guidance of the Food and Drug Administration, Health Canada and other health authorities, to suspend aprotinin from worldwide markets.8
Although the hemostatic derangements of CPB are more significant in pediatric patients, aprotinin’s efficacy and safety profiles in this patient population are less clear. Several prospective studies examining children undergoing re-operative cardiac surgical procedures found aprotinin to be effective in attenuating post-bypass coagulopathies and decreasing blood product exposure.3,9,10 A meta-analysis of 578 surgical procedures by Arnold et al. concluded that aprotinin was effective in reducing the amount of red blood cells or whole blood transfusions received by children undergoing CPB.11 This effect remained significant even in children undergoing primary sternotomy and in children weighing <10 kgs. However, while many studies in the pediatric population have mentioned the lack or rarity of complications associated with aprotinin, none of them has directly evaluated its safety. The ability to confidently assess the safety profile of aprotinin in pediatric patients has been hindered by several major obstacles: a heterogeneous patient population in terms of age and size, a dramatic variation in the types of procedures required by children with congenital heart disease, a magnitude of institutional differences governing the management of CPB and an inconsistency in aprotinin dosing regimes.
In light of the recent evidence in adult cardiac patients associating the intraoperative use of aprotinin with renal dysfunction, we examined a homogeneous group of pediatric patients who received a consistent dose of aprotinin and sought to determine if an association between aprotinin administration and postoperative renal dysfunction could be identified. To accomplish this, we performed a retrospective analysis of neonates who had undergone complex congenital heart surgery during a 2-yr period. Our objective was to determine the safety profile of aprotinin in these neonates in terms of 72-h postoperative renal dysfunction and the need for postoperative dialysis. We also examined the incidence of postoperative thrombosis and in-hospital mortality.
With IRB approval, we studied all neonates undergoing complex congenital heart surgery requiring CPB from January 1, 2005 through February 28, 2007. Chart reviews were performed by three of the authors of this manuscript who standardized among themselves precise definitions of the collected variables. Two hundred five neonates were included in the study. Four neonates were excluded due to death from severe myocardial dysfunction within the first 24 h of the operation. Of these four, only one received aprotinin. A fifth was excluded due to a second major operation on the first postoperative day. Two underwent two cardiac surgical procedures during their hospitalization; only the first procedure was included in the database. Two hundred neonates undergoing 200 cardiac surgical procedures were included in the final analysis. Neonates were divided into two groups: those who received aprotinin intraoperatively (aprotinin group) and those who did not (no aprotinin group).
Chart reviews consisted of preoperative, intraoperative and postoperative data collection. Preoperative data included patient demographics: age, weight, prematurity (defined as <36 wks gestation12), and presence of a major noncardiac structural anomaly. Cardiac diagnosis and surgical procedure were recorded. Neonates were assigned a Risk Adjustment for Congenital Heart Surgery (RACHS-1) score.12 We were unable to assign a RACHS-1 score to 2 patients who underwent cardiac transplantation because there is no RACHS-1 score for this procedure. Neonates with a combination of cardiac surgical procedures were assigned the RACHS-1 score of the highest risk procedure and, when applicable, the presence of a combination procedure was treated as an independent variable. Preoperative, baseline Cr levels were also recorded. Intraoperative data included the use of aprotinin, CPB time, aortic cross-clamp time, regional perfusion time, deep hypothermic circulatory arrest time and the lowest temperature achieved during CPB. Postoperative data assessed specific outcomes. Postoperative Cr levels were recorded at 24 and 72 h. Renal dysfunction was defined as a more than 50% increase in the preoperative Cr concentration at 72 h postoperatively.6,13 We recorded the highest temperature during the first 24 h postoperatively, time to tracheal extubation and duration of intensive care unit (ICU) stay. Notations were made of specific documented postoperative events: dialysis (peritoneal or hemodialysis), thrombosis documented by cardiac catheterization or echocardiography and mortality before hospital discharge.
Neonates received aprotinin during their surgical procedure as directed by the attending anesthesiologist. In all patients receiving aprotinin, the dose was based on a body surface area-calculated regime: 240 mg/m2 as a load to the patient and as a load to the CPB pump prime and an infusion of 56 mg/m2/h throughout the duration of the surgical procedure.14,15 For the average neonate in our patient population, this is the equivalent of 118,000 KIU/kg as a load to the patient and as a load to the CPB pump prime and an infusion of 27,500 KIU/kg/h throughout the duration of the surgical procedure. Anticoagulation was achieved by the administration of 400 U/kg of porcine heparin. Kaolin-activated activated clotting time (ACT) values exceeding 480 s were confirmed prior to the initiation of CPB. Additional heparin was administered as necessary during CPB to maintain an ACT > 480 s and, in neonates receiving aprotinin, an additional 100 U/kg of heparin was administered every hour during CPB as per institutional protocol. Nonpulsatile hypothermic CPB was performed using a nonheparin-coated system and a Terumo RX-05 hollow-fiber membrane oxygenator (Terumo Cardiovascular Systems, Ann Arbor, MI). COBE SMArt neonatal circuits (Sorin Group, Arvada, CO) containing a 300 mL priming volume with an additional 1000 U of heparin were used in all cases. Packed red blood cells were added to the pump prime and during CPB as needed to achieve a hematocrit of 30%. Modified ultrafiltration (MUF) was performed immediately after separation from CPB at the surgeon’s discretion. Protamine (4 mg/kg) was used to neutralize heparin upon completion of MUF. After confirmation of heparin neutralization by ACT, persistent bleeding was treated with transfusion of platelets followed by cryoprecipitate as deemed necessary by the attending anesthesiologist. In the ICU, the attending intensivist was the primary director of postoperative care.
All statistical analyses were performed using SAS (v.9.1) software (SAS Institute, Cary, NC) and statistical significance was assessed at the 0.05 level. Collected data were summarized descriptively and compared between the aprotinin and no aprotinin groups. For continuous measurements, means were calculated and tested for equality between groups using Student’s t-test. Categorical outcomes were summarized with percentages and compared using the χ2 test of independence or Fisher’s exact test when small expected values were encountered. Preoperative 24-h and 72-h postoperative serum Cr levels were analyzed using a longitudinal Mixed Model. We used this model specifically to adjust for both the correlated nature of the repeated measurements taken on each neonate and the three missing 72-h Cr values. This procedure was performed using SAS PROC MIXED with an “unstructured” covariance matrix.
The primary outcome of postoperative renal dysfunction as well as other measured outcomes (highest temperature during the first 24 h postoperatively, duration of mechanical ventilation, length of ICU stay, dialysis, thrombosis, and in-hospital mortality) were also summarized and compared between both groups. Due to the skewed nature of the data, continuous outcomes were summarized by medians and compared using the Wilcoxon’s rank-sum test. Categorical outcomes were summarized by percentages and compared using either the χ2 test of independence or Fisher’s exact test as described above.
To assess univariate associations with postoperative renal dysfunction, all variables were compared between neonates who developed renal dysfunction and those who did not, using the same summary statistics and statistical tests as described above. A scatterplot suggested that renal dysfunction only occurred in neonates with CPB duration longer than 100 min; thus the same variables were assessed in this subgroup using similar methods.
Stepwise logistic regression was performed for all variables that were either significantly associated with aprotinin administration or significantly associated with renal dysfunction for all 200 neonates. To confirm that these results were not influenced by a model overloaded with too many variables for the small number of renal dysfunction events, we next performed a multivariate logistic regression limiting the independent variables to only those found by univariate analysis to be associated with renal dysfunction. The administration of aprotinin was also included in this assessment as an additional independent variable. Adjusted odds ratios for renal dysfunction, as well as their 95% confidence intervals, were calculated. Model fit was verified by the Hosmer-Lemeshow test and model predictive ability was quantified by the area under the curve.
Two hundred consecutive neonates presenting for complex congenital heart surgery requiring CPB were included in this retrospective analysis. One hundred fifty-six neonates were in the aprotinin group and 44 were in the no aprotinin group. Patient demographics and CPB data are shown in Table 1. The aprotinin group had a statistically significantly larger percentage of males and a larger percentage of neonates undergoing RACHS-1 score 4 procedures. Additionally, neonates in the aprotinin group were exposed to statistically significantly longer CPB and aortic cross-clamp times, lower temperatures on CPB and more frequent use of deep hypothermic circulatory arrest and MUF. The no aprotinin group had a larger percentage of neonates undergoing RACHS-1 score 2 procedures. The cardiac surgical procedures of each neonate are delineated in Table 2.
Adjusted mean serum Cr levels preoperatively and at 24 and 72 h postoperatively were calculated for each group (Table 3). Mean preoperative Cr levels were similar between neonates in the aprotinin group and those in the no aprotinin group. At 24 and 72 h postoperatively, mean Cr levels were also similar between the two groups. The change in Cr over time was highly significant (P < 0.0001) and, again, similar between the two groups (P = 0.21). At 24 h postoperatively, the mean changes in Cr levels for both groups were statistically elevated significantly from their respective preoperative levels. At 72 h postoperatively, Cr levels had decreased significantly from 24 h levels, but still remained significantly elevated from baseline values in both groups. The mean increase in Cr from baseline to 72 h postoperatively was 0.14 mg/dL (95% CI: +0.09, +0.18) in the aprotinin group vs 0.07 mg/dL (95% CI: +0.01, +0.14) in the no aprotinin group (P = 0.15).
A larger percentage of neonates in the aprotinin group reached Cr levels that were used to define renal dysfunction at 72 h postoperatively than neonates in the no aprotinin group, though this difference failed to reach statistical significance (20.1% aprotinin vs 9.3% no aprotinin; P = 0.12). There were no other statistically significant differences between the two groups with regard to measures of patient outcome [values expressed as median (Q1, Q3), highest temperature within the first 24 h postoperatively: 37.2°C (37.0°C, 37.5°C) aprotinin vs 37.2°C (37.0°C, 37.5°C) no aprotinin, P = 0.88; duration of mechanical ventilation: 3.1 (2.6, 5.3) d aprotinin vs 3.8 (1.7, 5.7) d no aprotinin, P = 0.85; length of ICU stay: 7 (5, 10) days aprotinin vs 6 (4, 11) days no aprotinin, P = 0.81]. Additionally, other documented adverse events also showed no differences between the two groups [dialysis: 5.1% aprotinin vs 6.8% no aprotinin, P = 0.71; thrombosis: 4.5% aprotinin vs 4.6% no aprotinin, P = 1.00; in-hospital mortality: 9.0% aprotinin vs 13.6% no aprotinin, P = 0.40]. Inclusion of the four neonates, who died postoperatively within the first 24 h changes in-hospital mortality percentages to 9.6% for the aprotinin group and 19.2% for the no aprotinin group (P = 0.07).
Patient demographics and CPB data for the 35 neonates who developed postoperative renal dysfunction are shown in Table 4A. We observed statistically significant differences in these neonates with regards to age, CPB time, aortic cross-clamp time and lowest temperature. Since evidence of multicollinearity could be demonstrated among the CPB variables, we used CPB time as an independent variable and plotted the percent change in Cr from baseline to 72 h after surgery against CPB time (Fig. 1). We found that postoperative renal dysfunction only occurred in those neonates with a CPB time of longer than 100 min. Neonates with a CPB time of <100 min did not develop postoperative renal dysfunction. Table 4B compares those neonates with a CPB time of more than 100 min who did not develop renal dysfunction with those neonates who did. Only age and CPB time were significantly different between these two groups. Again, aprotinin was not found to be significant. Furthermore, within this subgroup, the percent of neonates in the aprotinin group who developed renal dysfunction was not statistically significantly different from the percent in the no aprotinin group (Table 5).
Stepwise logistic regression was performed to determine the impact on renal dysfunction of each of the variables whose mean values differed between neonates who did or did not receive aprotinin (Table 1) and between neonates who did or did not develop postoperative renal dysfunction (Table 4A). From this analysis, the only significant predictors of postoperative renal dysfunction that emerged were CPB time and age. Next, a multivariate logistic regression analysis, again including all neonates in the study and using postoperative renal dysfunction as the dependent variable, was performed using only CPB time, age and the use of aprotinin as the independent variables. CPB time remained the most significant predictor (P = 0.005), age was no longer significant (P = 0.08) and aprotinin administration continued to be insignificant (P = 0.45). The adjusted odds ratio of renal dysfunction for CPB time was 1.16 (95% CI: +1.05, +1.29) indicating a 16% increase in the odds of renal dysfunction for every 10 min of CPB time over 100 min. The predictive ability of the model was fair with an estimated area under the curve of 0.71.
In this investigation, we found that the mean serum Cr level of neonates undergoing complex congenital heart surgery requiring CPB at our institution was elevated at both 24 and 72 h postoperatively. These perioperative changes in serum Cr were highly significant and occurred regardless of whether or not aprotinin was administered. The greatest increase in serum Cr was found at 24 h postoperatively and was similar between the aprotinin and no aprotinin groups. Subsequently, at 72 h postoperatively, mean Cr levels in both groups began to trend back toward preoperative levels, but still remained significantly higher than baseline. The percentage of neonates reaching our criterion for renal dysfunction at 72 h postoperatively in the aprotinin group was larger than the percentage in the no aprotinin group but this difference failed to reach statistical significance. We also found that the aprotinin group had significantly longer CPB and aortic cross-clamp times and were cooled to significantly lower temperatures during CPB. More neonates in the aprotinin group underwent deep hypothermic circulatory arrest and MUF. A stepwise logistic regression examining the differing variables between the two groups identified CPB time and age as the most significant predictors of 72-h postoperative renal dysfunction. Indeed, a CPB time of 100 min appeared to be a critical marker for the development of postoperative renal dysfunction in this study. Neonates with a CPB time of <100 min were not at risk for the development of postoperative renal dysfunction, whereas neonates with a CPB time of more than 100 min were. This effect remained true regardless of the use of aprotinin. Therefore, we conclude that the complexity of the surgery, i.e., those procedures with longest CPB times, had the greatest impact on predicting the occurrence of postoperative renal dysfunction.
Aprotinin administration is associated with a dose-dependent attenuation of fibrinolytic activity in both adults and children during CPB,16–18 and thus the potential for postoperative thrombosis has always been a concern. This concern, however, has not consistently been supported by the pediatric literature. In a meta-analysis of 12 pediatric studies, seven of the studies reported adverse events and observed no increase in thrombotic complications in aprotinin-treated children.11 Conversely, in a study examining the safety of aprotinin in 60 children weighing <15 kg, Carrel et al. reported seven thrombotic complications with five of them occurring in aprotinin-treated children.19 These authors, however, did not confirm by statistical analysis whether this incidence was significantly different between treatment groups. In slightly older children, Penkoske et al. also reported three thrombotic complications in their aprotinin-treated children, but again did not perform statistical testing to demonstrate significant between-group differences.20 Although some of these investigations included neonates, the numbers were small and the focus was not the neonatal population. Our study found that the administration of aprotinin to neonates did not statistically significantly increase the incidence of documented postoperative thrombotic complications. However, because this study was retrospective, we could only note those episodes of thrombosis that were accurately documented in the hospital chart.
In addition to its antifibrinolytic activity, aprotinin may have an attenuating effect on the inflammatory response to CPB. This effect could be especially beneficial to neonates who often exhibit a marked inflammatory reaction and capillary leak syndrome during CPB.21 Due to the retrospective nature of our study, we did not examine specific biomarkers of inflammation. However, clinical variables most likely to be affected by the inflammatory response, such as highest temperature in the first 24 h postoperatively,22,23 duration of mechanical ventilation and length of ICU stay, showed no statistically significant differences between the two groups, suggesting that any potential antiinflammatory effect of aprotinin was clinically negligible in the long run.
Most studies in the adult cardiac literature evaluating the safety of aprotinin’s postoperative effects have included thousands of patients, well above the number included in this study. We acknowledge that the sample size of this study is small and perhaps, in some instances, may be responsible for our inability to show a statistically significant difference between the two groups. However, our report does represent the largest collection to date of neonates from a single institution using relatively consistent clinical practices examining the safety profile of aprotinin in this high risk population. Nevertheless, because of the limited power of our study, we were unable to make definitive conclusions regarding the safety of aprotinin in neonates undergoing complex congenital cardiac surgery requiring CPB.
This study suffers from other limitations of a retrospective investigation. First, we had no control over the assignment of aprotinin between the two groups and, as a result, the aprotinin group is larger than the no aprotinin group. Additionally, 12 of the patients in the no aprotinin group received tranexamic acid. However, since the primary goal of this investigation was to examine the safety of aprotinin vs all other treatment options, we felt that including these 12 neonates in the no aprotinin group was valid. A second limitation is the difference between the two groups in regards to CPB variables. We believe that this indicates our institutional prejudice to use aprotinin in the more complex cases, thus making it difficult to definitively discern the effects of CPB time from the effects of aprotinin administration. Third, we were unable to control for the use of MUF after CPB. The use of MUF was determined by surgeon preference and, again, was used most in the more complex cases. Nonetheless, in the overall group and in the subgroup with a CPB time of longer than 100 min, MUF was not found to be a significant predictor of postoperative renal dysfunction. A final limitation of this investigation is its short follow-up period. Patient records were reviewed only until hospital discharge. Thus, we have no data to suggest whether there are long-term mortality differences between the two groups.
In conclusion, our data suggest that neonates undergoing complex congenital heart surgery requiring CPB at our institution had significant elevations in 24- and 72-h postoperative serum Cr levels. These elevations occurred regardless of whether or not aprotinin was administered intraoperatively. We found no statistically significant difference between the percent of neonates in each group who developed postoperative renal dysfunction, though one might speculate that this difference could become significant with a larger sample size. Stepwise logistic regression identified CPB time and age as significant predictors of postoperative renal dysfunction, and a second multivariable regression found CPB time to be the most significant predictor. Indeed, neonates exposed to more than 100 min of CPB were at greatest risk for the development of postoperative renal dysfunction. In the neonates receiving aprotinin, there was no increase in renal failure requiring dialysis, no increase in postoperative thrombotic complications and no increase in in-hospital mortality. Because of our relatively small sample size and the retrospective protocol of this study, the credibility of our results would be enhanced by a larger, prospective investigation of the effects of aprotinin in neonates undergoing complex congenital heart surgery requiring CPB. We hope that one day aprotinin will be available for such a study.
1. Kovesi T, Royston D. Pharmacological approaches to reducing allogeneic blood exposure. Vox Sang 2003;84:2–10
2. Levy JH, Pifarre R, Schaff HV, Horrow JC, Albus R, Spiess B, Rosengart TK, Murray J, Clark RE, Smith P, Nadel A, Bonney SL, Kleinfield R. A multicenter, double-blind, placebo-controlled trial of aprotinin for reducing blood loss and the requirement for donor-blood transfusion in patients undergoing repeat coronary artery bypass grafting. Circulation 1995; 92:2236–44
3. D’Errico CC, Shayevitz JR, Martindale SJ, Mosca RS, Bove EL. The efficacy and cost of aprotinin in children undergoing reoperative open heart surgery. Anesth Analg 1996;83:1193–9
4. Mossinger H, Dietrich W, Braun SL, Jochum M, Meisner H, Richter JA. High-dose aprotinin reduces activation of hemostasis, allogeneic blood requirement, and duration of postoperative ventilation in pediatric cardiac surgery. Ann Thorac Surg 2003;75:430–7
5. Mangano DT, Tudor IC, Dietzel C. The risk associated with aprotinin in cardiac surgery. N Engl J Med 2006;354:353–65
6. Karkouti K, Beattie WS, Dattilo KM, McCluskey SA, Ghannam M, Hamdy A, Wijeysundera DN, Fedorko L, Yau TM. A propensity score case-control comparison of aprotinin and tranexamic acid in high-transfusion-risk cardiac surgery. Transfusion 2006;46:327–38
7. Mangano DT, Miao Y, Vuylsteke A, Tudor IC, Juneja R, Filipescu D, Hoeft A, Fontes ML, Hillel Z, Ott E, Titov T, Dietzel C, Levin J. Mortality associated with aprotinin during 5 years following coronary artery bypass graft surgery. JAMA 2007;297:471–9
9. Davies MJ, Allen A, Kort H, Weerasena NA, Rocco D, Paul CL, Hunt BJ, Elliott MJ. Prospective, randomized, double-blind study of high-dose aprotinin in pediatric cardiac operations. Ann Thorac Surg 1997;63:497–503
10. Miller BE, Tosone SR, Tam VK, Kanter KR, Guzzetta NA, Bailey JM, Levy JH. Hematologic and economic impact of aprotinin in reoperative pediatric cardiac operations. Ann Thorac Surg 1998;66:535–41
11. Arnold DM, Fergusson DA, Chan AKC, Cook RJ, Fraser GA, Lim W, Blajchman MA, Cook DJ. Avoiding transfusions in children undergoing cardiac surgery: a meta-analysis of randomized trials of aprotinin. Anesth Analg 2006;102:731–7
12. Jenkins KJ, Gauvreau K, Newburger JW, Spray TL, Moller JH, Iezzoni LI. Consensus-based method for risk adjustment for surgery for congenital heart disease. J Thorac Cardiovasc Surg 2002;123:110–8
13. Bellomo R, Ronco C, Kellum JA, Mehta RL, Palevsky P; ADQI workgroup. Acute renal failure—definition, outcome measures, animal models, fluid therapy and information technology needs: the second international consensus conference of the Acute Dialysis Quality Initiative (ADQI) group. Crit Care 2004;8:R204–R12
14. Elliot MJ, Allen A. Aprotinin in paediatric cardiac surgery. Perfusion 1990;5:73–6
15. Costello JM, Backer CL, de Hoyos A, Binns HJ, Mavroudis C. Aprotinin reduces operative closure time and blood product use after pediatric bypass. Ann Thorac Surg 2003;75:1261–6
16. Marx G, Pokar H, Reuter H, Doering V, Tilsner V. The effects of aprotinin on hemostatic function during cardiac surgery. J Cardiothorac Vasc Anesth 1991;5:467–74
17. Lu H, Soria C, Commin PL, Soria J, Piwnica A, Schumann F, Regnier O, Legrand Y, Caen JP. Hemostasis in patients undergoing extracorporeal circulation: the effect of aprotinin (Trasylol). Thromb Haemost 1991;66:633–7
18. Dietrich W, Mossinger H, Spannagl M, Jochum M, Wendt P, Barankay A, Meisner H, Richter JA. Hemostatic activation during cardiopulmonary bypass with different aprotinin dosages in pediatric patients having cardiac operations. J Thorac Cardiovasc Surg 1993;105:712–20
19. Carrel TP, Schwanda M, Vogt PR, Turina MI. Aprotinin in pediatric cardiac operations: a benefit in complex malformations and with high-dose regimen only. Ann Thorac Surg 1998;66:153–8
20. Penkoske PA, Entwistle LM, Marchak BE, Seal RF, Gibb W. Aprotinin in children undergoing repair of congenital heart defects. Ann Thorac Surg 1995;60:S529–S32
21. D’Errico CC, Munro HM, Bove EL. Pro: the routine use of aprotinin during pediatric cardiac surgery is a benefit. J Cardiothorac Vasc Anesth 1999;13:782–4
22. Mitchell JD, Grocott HP, Phillips-Bute B, Mathew JP, Newman MF, Bar-Yosef S. Cytokine secretion after cardiac surgery and its relationship to postoperative fever. Cytokine 2007;38:37–42
© 2009 International Anesthesia Research Society
23. Madhok AB, Ojamaa K, Haridas V, Parnell VA, Pahwa S, Chowdhury D. Cytokine response in children undergoing surgery for congenital heart disease. Pediatr Cardiol 2006; 27:408–13