| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ann Thorac Surg 2007;84:1320-1325
© 2007 The Society of Thoracic Surgeons
a Department of Pediatrics, University of Rochester Medical Center, Rochester, New York
b Department of Surgery, University of Rochester Medical Center, Rochester, New York
c Department of Biostatistics and Computational Biology, University of Rochester Medical Center, Rochester, New York
d Department of Anesthesiology, University of Rochester Medical Center, Rochester, New York
Accepted for publication May 11, 2007.
* Address correspondence to Dr Lerner, 601 Elmwood Ave, URMC Box 777, Rochester, NY 14642 (Email: norma_lerner{at}urmc.rochester.edu).
 |
Abstract |
|---|
 Top Abstract IntroductionPatients and MethodsResultsCommentAcknowledgmentsReferences |
|---|
Methods: Neonates (<30 days) requiring initial surgical palliation were studied. All subjects received aspirin from postoperative day 1 onward. Thromboses were diagnosed by serial transthoracic echocardiograms, vascular imaging, and interstage cardiac catheterizations according to predefined criteria.
Results: Twenty-two neonates, age 1 to 11 days (mean 4 ± 2.5) were studied. Follow-up ranged from three hours to 18 months (median, 212 days). Eight infants died. Four of the 14 subjects who survived (28%), and one of the eight who died (12.5%), had evidence of thrombosis identified over a range of four hours to nine months postoperatively (median 14 days). When compared with reference values established in healthy children, preoperative subject hematocrit (Hct), platelet count, factors II, V, VII, VIII, and X, antithrombin, protein C, and soluble CD40 ligand measures were significantly lower, and the prothrombin time and partial thromboplastin time were significantly higher. Postoperative C reactive protein (CRP) was significantly higher, and Hct and platelet count significantly lower, than preoperative values. Thrombotic events were significantly related to high preoperative CRP (p = 0.02).
Conclusion: Thrombotic complications occur frequently in neonates undergoing initial palliative surgery, suggesting that aspirin therapy alone may constitute inadequate protection. Elevated preoperative CRP appears to be associated with increased thrombotic risk.
|
Introduction |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentAcknowledgmentsReferences |
|---|
Less information is available regarding the incidence of thrombosis associated with initial (neonatal) palliative surgeries, including the modified Norwood procedure, and those surgeries to create systemic to pulmonary blood flow (Blalock-Taussig [BT] shunts and right ventricle [RV] to pulmonary artery [PA] conduits). In one retrospective study of neonates having initial surgical palliation, mortality before BDG was approximately 12%, and in patients for whom autopsies were available, 33% of deaths were attributed to thrombosis [4]. In the sole prospective study of subjects undergoing initial palliation, 45% of patients died, and 25% developed thrombosis [5]. No trials have prospectively investigated thrombotic risk factors in neonates prior to initial palliative repairs.
While not directly addressed in children undergoing cardiac surgery, there is substantial evidence that inflammation has a pathogenic role in vascular disease. Studies have confirmed the utility of wide-range C-reactive protein (wrCRP) for assessing the risk of vascular disease [6]. Fibrinogen and factor VIII are elevated in adults with cardiovascular disease and offer risk prediction for future cardiovascular events [7]. The CD40 ligand (CD40L; formally CD154), member of the tumor necrosis family, is both prothrombotic and proinflammatory. Soluble CD40L (sCD40L) is primarily platelet-derived and is significantly increased in adult acute coronary syndromes [8] and after CPB [9]. Elevated levels are also associated with cardiovascular event risk in healthy women [10].
The primary objective of this study was to prospectively evaluate the incidence of thrombosis in neonates undergoing initial surgical palliation for complex congenital heart disease, specifically variations of single ventricle physiology. Thrombi developing in the context of initial palliation procedures may lead to death and, by occluding major vessels, may complicate postoperative medical management and later cardiac catheterization and operative procedures. Those that obstruct the cavopulmonary vascular bed can cause pulmonary hypertension that may preclude later operative repairs.
The secondary objective of the study was to determine whether particular laboratory test abnormalities, reflecting aberrations of coagulation and (or) inflammation, might serve as markers of increased thromboembolic risk. Inflammatory markers were included based upon the hypothesis that, as found in adult vascular disease, inflammation may be pathogenic in children with complex congenital heart disease.
|
Patients and Methods |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentAcknowledgmentsReferences |
|---|
Subjects
Inclusion criteria were the following: (1) age less than 30 days; (2) congenital cardiac disease with single ventricle physiology requiring surgical palliation to create systemic to pulmonary blood flow, regardless of disease severity or associated conditions; and (3) informed consent.
All subjects were hemodynamically stable before proceeding to surgery; none were palliated emergently. No subjects required inotropic support preoperatively and few were mechanically ventilated. Neonates with ductal-dependent lesions were maintained preoperatively on prostaglandin infusions. None of the subjects were premature or small for gestational age.
Twenty-five eligible neonates were identified between March 2005 and October 2006. Three families were unavailable for consent. Twenty-two (88%) were approached for study enrollment; all agreed to participate. Properly witnessed informed written consent was obtained. No subject withdrew, and none was lost to follow-up.
Clinical Management
Subjects requiring CPB received methylprednisolone (3 mg/kg, intravenously) the night prior to surgery. When appropriate, standard surgical and CPB techniques, with core cooling to 20°C when necessary, were employed. Deep hypothermic circulatory arrest was limited by the use of regional antegrade cerebral perfusion and bicaval cannulation whenever possible. Fresh (<14 days), washed, leukoreduced irradiated packed red blood cells were used to prime the CPB circuit. Myocardial protection was accomplished by antegrade cold blood cardioplegia. Aprotinin infusion was administered in all but two cases where aminocaproic acid was given. Heparin was neutralized by protamine (0.7 to 0.8 mg per 100 units of bolused heparin) with additional protamine if the activated clotting time did not normalize.
After operation, subjects received inotropes and (or) inodilators as needed to maintain adequate cardiac output. Central venous lines (CVL) were removed once the subject was successfully extubated and off inotropes-pressors, typically postoperative day (POD) 3 to 7. Arterial lines were maintained with heparinized saline, as were CVLs when the central venous pressure was continuously monitored.
Beginning POD 1 all subjects received half of an aspirin suppository (40.5 mg) daily (10 to 15 mg/kg/dose). Subjects transitioned to oral aspirin when able to take by mouth (typically POD 5 to 7). No additional form of anticoagulation was utilized unless a thrombotic complication was identified, at which time individualized treatment was initiated.
Laboratory Studies
Baseline preoperative studies were drawn the day of surgery, prior to CPB, when each subject was hemodynamically stable and free of known infection. Postoperative day 3 was selected for follow-up studies as subjects would have returned closer to "steady-state" function while still having CVL access.
With the exception of plasminogen and plasminogen activator inhibitor (PAI-1), all laboratory tests were analyzed at the URMC. The white blood count, Hct, and platelet counts were determined by argon-ion laser light scatter (MAPSS technology) and focused flow impedance with two-color fluorescent flow cytometry (AbbottCell Dyn 4000 Hematology Analyzer; Abbott Laboratories, Abbott Park, IL). The wrCRP was analyzed by immunoturbidimetric assay (ADVIA 2400 Chemistry System-Bayer ADVIA kit; Bayer Healthcare, Tarrytown, NY). Coagulation factors were measured according to manufacturer specifications and standard laboratory methods [11]. Antithrombin and protein C activity were performed using a chromogenic method with antithrombin and protein C reagents. Protein S activity was analyzed using dual channel photo-optical clot detection (MDA II analyzer and protein S reagent; bioMerieux Inc, Hazelwood, MO). The prothrombin time, partial thromboplastin time, fibrinogen, and factors were determined by optical clot detection (MDA II analyzer). Simplastin, platelin, fibroquick, and appropriate factor-deficient plasma reagents were employed. The sCD40L levels were determined using an enzyme-linked immunosorbent assay protocol that captures monomeric and multimeric protein forms, with a method described previously [12].
Molecular diagnostic testing for factor V Leiden (G1691A) and prothrombin 20210A mutations were performed by polymerase chain reaction followed by hybridization with allele specific DNA fluorescent probes. Plasminogen and PAI-1 were determined at the Associated Regional and University Pathologists clinical laboratories (Salt Lake City, Utah). Percent plasminogen activity was analyzed by chromogenic assay, and PAI-1 activity analyzed by bioimmunoassay.
Surveillance Studies
A transthoracic echocardiogram (TTE) was performed at diagnosis, on the day of discharge, and for any concern regarding cardiac function or development of thrombosis. During each TTE, extended views of the vasculature were pursued to identify intracardiac thrombi, thrombi in any shunt-conduit, or any arterial or venous thrombus in a "large" vessel. Large vessels were defined as femoral arteries-veins, iliac arteries-veins, subclavian arteries-veins, the internal jugular veins, and the inferior and superior vena cava.
Cardiac catheterizations performed prior to second operative repair provided additional information regarding intracardiac-conduit thrombi and vessel patency. Any thrombus causing alteration in blood flow or obstruction was considered significant. Additional imaging (venous and [or] abdominal ultrasound or TTE) was performed for clinical suspicion of thrombus or vascular obstruction.
Statistical Analysis
Paired ttests (2-tailed) and Wilcoxon signed rank tests were used to compare preoperative and postoperative means or medians. When established pediatric norms with SD and sample sizes were available, unpaired, 2-tailed t tests or rank tests were performed to compare subject and reference means or medians.
To analyze the effect of laboratory test levels on the categoric variable of thrombosis, the study sample was divided into two groups; the first comprised of those subjects with values less than the 90th percentile of normal reference values, and the second of those subjects with levels greater than or equal to that value. For studies in which a depressed value would be anticipated in the context of thrombosis, the first group was comprised of subjects with values equal to or less than the tenth percentile and the second of those with values greater than the tenth percentile. The relationship between test group and thrombus formation was assessed using the Fisher exact test.
Correlation coefficients examined the relationships between platelet count and sCD40L. For all analyses, a p value of 0.05 or less was considered statistically significant. Analyses were performed using MedCalc version 9.2 for Windows (Mariakerke, Belgium).
|
Results |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentAcknowledgmentsReferences |
|---|
|
Clinical Outcomes
Five subjects had evidence of thrombotic complications identified four hours to nine months after surgery (median, 14 days). Details are summarized in Table 2.
|
Laboratory Studies
Test results are presented in Table 3. Twenty-two samples were obtained for each variable preoperatively, and at least 15 samples postoperatively (due to preceding deaths).
|
All laboratory results were compared with normal established pediatric reference ranges as shown in Table 3 [13–18]. There was no correlation between platelet number and serum or plasma sCD40L; preoperative: platelet count versus serum sCD40L (n = 24, r = 0.08 [95% confidence interval {CI} –0.33 to 0.47], p= 0.70); platelet count versus plasma sCD40L (n = 24, r = –0.07 [95% CI –0.46 to 0.34], p = 0.74); postoperative: platelet count versus serum sCD40L (n = 21, r = –0.12 [95% CI –0.53 to 0.32], p = 0.60); platelet count versus plasma sCD40L (n = 21, r = –0.21 [95% CI –0.59 to 0.24], p = 0.36).
While plasminogen activity was substantially lower than reported normal adult ranges, the mean and standard deviation for this test were not available, and significance could not be established. Preoperative wrCRP reflected established ranges and postoperative PAI-1 and wrCRP were higher than reference values, but a lack of information precluded testing for statistical significance (Table 3).
The association between all laboratory values and the clinical outcome of thrombosis was assessed. The percentage of patients who developed thrombosis was significantly related only to wrCRP group (Fisher exact test, p = 0.02; Fig 1B).
|
|
Comment |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentAcknowledgmentsReferences |
|---|
Thrombi developed despite early administration of aspirin, suggesting that more aggressive anticoagulation may be warranted. In adults, where aspirin is used effectively for primary and secondary prevention of vascular disease [19], the antiplatelet effect of aspirin appears to vary. "Aspirin resistance," assessed by platelet aggregation testing or urinary thromboxane B2 levels, predicts adverse clinical outcomes [20–22]. Similar studies regarding aspirin in neonates with cardiac disease are needed. Because no patient in this series had bleeding complications, addition of another antiplatelet agent (ie, clopidogrel) [23, 24] or anticoagulant (heparin or warfarin) might be safely pursued. No prospective trials compare alternative postoperative anticoagulation strategies in patients undergoing initial palliation and approaches vary from institution to institution. A randomized controlled trial to compare postoperative regimens would provide additional information in this regard.
Compared with age-appropriate norms, mean values for natural anticoagulants, antithrombin, and protein C were significantly depressed. While this would imply a prothrombotic tendency, these aberrations were balanced by significant depressions in platelet number and factors VII, VIII, and X, all of which predispose to bleeding. The lower plasminogen and higher PAI-I levels, if significant, would contribute to decreased fibrinolytic capabilities in this population. It is interesting that all laboratory abnormalities appear not only to predate the later cavopulmonary connection, but to exist even earlier on, prior to surgery. This suggests an intrinsic coagulopathy, unrelated to the operative procedure.
Given its association with inflammation and thrombosis, preoperative and postoperative serum sCD40L and postoperative plasma sCD40L were, surprisingly, significantly lower than normal age-appropriate means. Similarly, when adults with stable cardiovascular disease were compared with those with unstable disease, serum sCD40L was found, again paradoxically, to be significantly lower in the unstable group [25]. There was no readily available explanation. Because 95% of circulating sCD40L is derived from platelets [26], our subjects’ depressed sCD40L might simply reflect lower platelet numbers (247 ± 104 vs 343 ± 72) than referenced for this age. There was no correlation, however, between our subjects’ sCD40L level and platelet count, although one cannot completely rule out such a relationship, given the small sample size and wide confidence intervals. Such a finding would not be unusual because a lack of correlation has previously been documented in children and adults [25, 27]. Alternatively, lower sCD40L levels may reflect that platelets in children with single ventricle physiology are, in a sense, "spent" and hypoactive compared with those from healthy infants.
Most importantly, elevated preoperative wrCRP appears to identify a subset of patients at increased risk for thrombotic events. Certainly the presence of newly introduced prosthetic material, central lines, hemodynamic instability, and intercurrent septic episodes predisposed these neonates to thrombosis. Nonetheless, other subjects with similar risk factors did not develop thromboses. An inflammatory state, reflected in an elevated CRP, may have provided the additional stimulus necessary to sustain thrombosis.
This study has limitations. The "true" incidence of thrombosis may not be reliably represented given the high number of early deaths. In addition, the sample size is small and the study is underpowered to prove the predictive value of tested risk factors. Larger groups of subjects will be needed to reliably explore some of our hypotheses. The ability of wrCRP to identify at risk patients will also need to be validated in a larger population.
|
Acknowledgments |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentAcknowledgmentsReferences |
|---|
|
References |
|---|
|
TopAbstractIntroductionPatients and MethodsResultsCommentAcknowledgmentsReferences |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
