The evaluation and management of postnatal thromboses

Transcription

1 STATE-OF-THE-ART The evaluation and management of postnatal thromboses (2009) 29, r 2009 Nature Publishing Group All rights reserved /09 $32 Division of Neonatology, Department of Pediatrics, University of Florida College of Medicine, Gainesville, FL, USA In the pediatric population, neonates have the highest risk for thromboembolism (TE), most likely due to the frequent use of intravascular catheters. This increased risk is attributed to multiple risk factors. Randomized clinical trials dealing with management of postnatal thromboses do not exist, thus, opinions differ regarding optimal diagnostic and therapeutic interventions. This review begins with an actual case study illustrating the complexity and severity of these types of cases, and then evaluates the neonatal hemostatic system with discussion of the common sites of postnatal thrombosis, perinatal and prothrombotic risk factors, and potential treatment options. A proposed step-wise evaluation of neonates with symptomatic postnatal thromboses will be suggested, as well as future research and registry directions. Owing to the complexity of ischemic perinatal stroke, this topic will not be reviewed. (2009) 29, ; doi: /jp ; published online 12 March 2009 Keywords: postnatal thrombosis; anticoagulation; thrombolysis; thromboembolism Case study A full-term male infant born through spontaneous vaginal delivery was admitted to the neonatal intensive care unit (NICU) shortly after birth due to a history of low APGAR scores and respiratory distress requiring mechanical ventilation. The pregnancy was complicated by gestational diabetes treated with glyburide and chorioamnionitis was evident during labor. The patient was started on ampicillin and gentamicin and appropriate fluid resuscitation was initiated. The patient improved over the next few days and was able to be weaned off of ventilator support. The patient had an umbilical arterial catheter (UAC) placed after birth that was removed on day 3 of life. On day 4 of life the patient s legs appeared pale. Physical examination demonstrated cool lower extremities and markedly Correspondence: Dr MA Saxonhouse, Division of Neonatology, Department of Pediatrics, University of Florida College of Medicine, PO Box , 1600 SW Archer Road, Gainesville, FL 32610, USA. saxonma@peds.ufl.edu Received 18 August 2008; revised 6 January 2009; accepted 21 January 2009; published online 12 March 2009 decreased pulses in the feet. Over the previous 24 h, urine output declined significantly. Ultrasound (US) examination of the heart, major vessels and abdomen demonstrated a large thrombus in the descending aorta with very minimal flow to the renal and iliac arteries. A literature search did not reveal any level I management guidelines for thrombosis in neonatal patients, and consideration is now given to surgical thrombectomy, anticoagulation therapy or fibrinolytic therapy. Incidence of neonatal thromboembolic disease Although neonates have the highest risk of thromboembolism (TE) in the pediatric population, the incidence of postnatal TE varies due to the types of thromboses that were reported and how aggressively centers screened for thromboses. 1 Data from the three international registries are displayed in Table 1. All three registries observed that thromboses occurred in both term and preterm infants and affected male and female infants equally, other than renal vein thrombosis (RVT) that affected more male neonates. 2 4 These registries also demonstrated that approximately 90% of venous thromboses in neonates were associated with central venous catheters (CVLs). 2 4 The recurrence rate of TE following symptomatic neonatal events ranges from 3.3 to 7%. 5 The international registries are the first step toward increasing recognition of this problem. Still, improvements need to be made, especially on how to screen effectively neonates with thrombi for prothrombotic disorders. There are little data available with respect to the presence of inherited prothrombotic risk factors in specific pediatric populations other than Jews and Caucasians. 6 The neonatal hemostatic system The neonatal procoagulant, anticoagulant and fibrinolytic systems are unique and differ from those of children and adults. 7 Although many of these unique qualities place a neonate in a relative prothrombotic state, they tend to be balanced by other factors that prevent a term or relatively well premature infant from experiencing spontaneous thromboses (Figure 1). 8,9 When this balance is disrupted (Table 2), the neonate is at high risk for developing thromboses. 2,14,15 Specifically, infection promotes clotting activation and catheters provide a center for thrombus formation. 16

2 468 Postnatal thromboses Table 1 Rates of neonatal thrombosis per registry 2 4 Registry German The Netherlands Canadian Rates of neonatal thromboembolism reported 5.1 symptomatic (all types of events) per live births 0.7 per live births for venous thrombosis 24 clinically apparent events (excluding stroke) per NICU admissions Abbreviation: NICU, neonatal intensive care unit. Hemorrhage Protein S, Protein C, Antithrombin, Heparin cofactor II, Plasminogen, Alpha 2 -antiplasmin, tissue factor pathway inhibitor α2-macroglobulin Thrombosis Factors II, VII, IX, and X Factors XI, XII Prekallikrein High-molecularweight kininogen Platelet function Factor VIII VWF Activity Larger RBC size Figure 1 Balance of the neonatal coagulation system. Coagulation properties that increase a neonate s risk for thrombosis include an increase in factor VIII concentrations (bold print) as well as enhanced whole blood measures of primary hemostasis and whole blood coagulation assays Additional factors than increase a neonate s risk for thrombosis include a decrease in the anticoagulant (protein C, protein S, antithrombin and heparin cofactor II) and fibrinolytic (plasminogen and a-antiplasmin) system proteins (light print). Properties that increase a neonate s risk for hemorrhage include an increase in a 2 macroglobulin concentrations (bold print) as well as a decrease in the vitamin-kdependent clotting factors II, VII, IX and X as well as contact factors (light print) Neonates also have a decrease in in vitro platelet aggregation and activation studies. 11 Thus, the increase in increase in specific procoagulant proteins and enhanced whole blood measures of primary hemostasis combined with the decrease in anticoagulant and fibrinolytic proteins is balanced by a decrease in other major procoagulant proteins preventing a well neonate from experiencing spontaneous thrombosis, despite being slightly shifted in a prothrombotic state With all the differences that the neonatal coagulation system demonstrates, if one simply reviews the pro- and anticoagulant factor quantities and properties, it appears that there is a slight shift toward thrombosis in the healthy term infant (Figure 1). This shift, though, tends to be balanced by unknown factors that prevent excessive bleeding and clotting in healthy neonates. Locations of postnatal thromboembolism, imaging modalities and management Postnatal TE events occur in the neonate in a number of locations, whether venous or arterial. A more detailed description of the common types of postnatal thromboses will be described. Table 2 Risk factors for the development of neonatal thromboembolism 1 7,10,11,14 50 During pregnancy During labor and delivery During neonatal period Infertility Emergent cesarean Central catheters section Oligohydramnios Severe bradycardia Congenital heart disease Maternal prothrombotic Instrumentation Sepsis/Necrotizing enterocolitis state (including autoimmunity) Preeclampsia Asphyxia Diabetes Respiratory distress syndrome Chorioamnionitis Dehydration Prolonged rupture of membranes Polycythemia Pulmonary hypertension Prothrombotic disorders (see Table 4) Surgery Extracorporeal membrane oxygenation Adapted with permission. 51 Arterial thromboses Iatrogenic arterial thromboses. The majority of postnatal arterial thromboses tend to be iatrogenic in nature 7 and are mainly related to complications from UACs, peripheral arterial catheters (PALs) and femoral arterial catheters. 52 UAC usage occurs in 10 to 64% of all NICU admissions. 53 A recent prospective study found that arterial thrombosis, as a complication of UAC use, was found in 23.4% of 47 babies who had a UAC placed. 17 The majority of thromboses were found in asymptomatic neonates after catheter removal using Doppler US. Of these patients, none required thrombolysis and none had symptoms or signs of acute vascular compromise. 17 Earlier reported complications of aortic thromboses secondary to UAC use include mesenteric ischemia, hypertension, renal dysfunction/failure, limb loss and congestive heart failure The most recent Cochrane review found that fewer clinical vascular complications were observed with high UAC position recommending that only high-positioned catheters be placed. 58 Although it seems that a shorter length of catheter in the aorta would lead to fewer complications, other factors such as turbulence of blood flow around the catheter tip and the effects on flow dynamics of flushing the catheter after blood drawing complicate this picture. 59 Other methods to reduce the risk of thrombosis, such as low-dose continuous heparin infusion at 1 U ml 1, were found to prolong catheter patency but not reduce the risk of thrombosis. 60 Regardless of whether a neonate has a UAC, PAL or femoral arterial line, suspicion or confirmation of an arterial thrombosis should prompt immediate removal of the catheter. The most

3 Postnatal thromboses 469 sensitive method for the diagnosis of arterial catheter-related TEs is contrast angiography. 18 However, due to its risk, it is generally reserved for older children and adults. Therefore, US is mainly used although it tends to underdiagnose clinically significant arterial TEs. 19 Specific types of treatment for arterial thromboses include close observation and supportive care, anticoagulation with unfractionated heparin (UFH) or low-molecular-weight heparin (LMWH), fibrinolytic therapy and/or surgery. The ultimate goal is the restoration of blood flow if there is evidence of limb or organ ischemia. These modalities will be discussed later in further detail. Spontaneous arterial thromboses. Spontaneous arterial thromboses, though extremely rare, are usually severe and mostly involve the aorta tending to mimic congenital heart disease. 3,4,6,61,62 Optimal imaging strategy is vital to make the correct diagnosis and to initiate the proper treatment to minimize end-organ complications. Decreased pulses or the presentation of a cool-mottled extremity should alert the clinician to a possible arterial thrombosis, especially if echocardiography does not determine evidence for aortic coarctation or other heart lesion that restricts left ventricular outflow. Treatment modalities will be discussed later. Venous thrombosis Catheter-related venous thrombosis (noncardiac). As care improves for our smallest and most critically ill neonates, so does our dependence on CVLs. Although percutaneous CVLs and other types of CVLs generally improve care, thrombosis, infection, catheter malfunction and death are major risks associated with their use. Symptomatic venous TE was reported with an incidence of 0.24 per admissions to NICUs, almost all of which were secondary to CVLs. 3 Mechanisms believed to contribute to CVL TE include damage to vessel walls, 63 disrupted blood flow, infusion of substances that damage endothelial cells 64 and thrombogenic catheter materials. 65 Umbilical venous catheters (UVCs) are routinely used in NICUs during the first week of life, and are usually removed during this time period to be replaced by peripheral CVLs if the infant is dependent on i.v. nutrition or antibiotics. One of the major concerns with UVC use is the risk of portal vein thrombosis. A recent study found that 21.4% of 28 babies with UVCs had thrombus formation, 4 of which were in the inferior vena cava (IVC) and 2 in the portal vein. 17 However, the venous thrombus was detected only after catheter removal and none of the patients needed treatment and all were asymptomatic. 17 Autopsy studies have estimated that 20 to 65% of infants who die with a UVC in situ have microscopic evidence of TE. 20,21,66 Signs and symptoms suggestive of a UVC thrombus include persistent infection of the line, persistent thrombocytopenia and dysfunction of the line. Attempts to minimize complications associated with UVCs focus on prompt removal after a short period of use. 67 The Centers for Disease Control and Prevention currently recommend use of UVCs be limited to 14 days. 68 A randomized trial compared long-term UVC use (up to 28 days) with short-term (7 to 10 days) use followed by peripheral CVL placement; investigators found that in the short-term group, 9% developed thrombi with 4% being clinically significant compared with 13 and 7%, respectively, in the long-term group. 69 Therefore, although rates were similar, limiting UVC use to as short a period as possible is recommended. 69 The gold standard for diagnosing venous TE is venogram, however, this modality is difficult in neonates, especially premature infants, due to poor venous access, lower glomerular filtration rates and rapid dispersion times due to increased cardiac output. 18 As a result, US is the most widely and safely used modality. However, false-negative results often occur when evaluating for venous thromboses in the upper central venous system due to obstruction of the distal subclavian veins by the clavicles and that compression of veins in a central location by the probe is not feasible due to the neonate s rib cage. 70 Long-term complications of venous TE include chronic venous obstruction, chylothorax, portal hypertension and embolism. 18,71,72 The recognition of children presenting with post-thrombotic syndrome who had a history of a neonatal thrombosis has been increasing. 72 Symptoms of this syndrome include chronic edema, skin discoloration, poor wound healing, impaired tissue viability, pain and limitation of limb function. 72 Management of a catheter-related venous TE requires prompt removal of the catheter. However, due to the risk for emboli, many clinicians advocate delay in CVL removal if a thrombus has been detected until 3 to 5 days after anticoagulant therapy has been given. 73 No clinical studies support this recommendation and thus this practice remains open to physician discretion. Anticoagulation treatment, if desired, should consist of either UFH or LMWH for 10 days to 3 months. 73 Thrombolytic therapy is recommended only for venous TE in neonates with major vessel occlusion that is causing a critical compromise of organs or limbs. 73 Intracardiac thromboses and thromboses in infants with complex congenital heart disease. The placement of CVLs (UVCs, percutaneous CVLs and surgically placed CVLs) in the right atrium is a subject of debate. 74,75 Placement of the catheter in the right atrium can lead to damage of the endocardium inducing either pericardial tamponade and/or the development of intracardiac thrombi. 76,77 The development of intracardiac vegetations secondary to line infections can expose the infant to prolonged infection and dissemination of septic emboli. 76 However, some centers have reported successful placement of CVLs in the right atrium if strict management guidelines are followed. 75 Thrombus formation in the right atrium is an emergent concern due to the risk for dissemination of emboli into the lungs or obstruction of the right pulmonary artery. 78 Owing to the risks

4 470 Postnatal thromboses associated with placement of CVLs in the right atrium, we recommend that this practice be avoided unless a center has strict guidelines and is comfortable with placement in the right atrium. Surgical approach to remove a right atrial thrombus is not feasible in the neonate, especially the premature infant, due to extreme technical difficulty and unacceptable risks. Multiple case reports have demonstrated success in using recombinant tissue-type plasminogen activator (rt-pa) for thrombolysis of catheter-related atrial thrombus in the neonate rt-pa has also been used successfully in the management of premature infants with infective endocarditis. 76 Specific details for dosing and management will be described later. Thrombosis is a common complication in infants and children undergoing repair for complex congenital heart disease. 82 These patients are not only dependent on CVLs for medications and nutrition, they are prone to recurrent infection. A retrospective review of neonates who underwent palliative repair for single ventricle physiology found that in patients who died and had autopsies available, 33% of deaths were attributable to thrombosis. 83 A more recent review evaluated 22 neonates who underwent palliative repair between 1 and 11 days. They found that five patients (23%) had evidence of thrombi over a range of 4 h to 9 months postoperatively. 82 The thrombi developed despite the early administration of aspirin and the study found that thrombotic events were significantly related to high preoperative C-reactive protein (CRP) values. Therefore, use of agents other that aspirin, especially if CRP values are elevated before initial palliation, may reduce the risk of thrombus formation. 82 However, larger trials are needed before definitive recommendations can be made. The diagnosis of either right atrial thrombus formation, intracardiac vegetations or thrombus formation in infants with single ventricle physiology is usually made by echocardiography. Signs suggestive of an atrial thrombus include new onset murmur, persistent sepsis, persistent thrombocytopenia and cardiac failure. Technicians and physicians skilled in the proper echocardiographic evaluation are important for management of these patients. Renal venous thrombosis. With an incidence of 0.5 per 1000 NICU admissions, RVT is the most common spontaneous venous TE in neonates. 3 A recent review of 271 neonatal cases demonstrated that 70% of cases were unilateral, with 64% of these involving the left kidney. 22 In accordance with earlier reviews, male neonates were more frequently affected. 22,84 Patients with RVT tend to present with at least one of the cardinal signs. A recent study demonstrated that approximately half of patients diagnosed with RVT had either macroscopic hematuria, a palpable abdominal mass and/or thrombocytopenia. 22 Retrospective studies that have evaluated neonates with RVT have demonstrated that 43 to 67% of these patients had at least one or more prothrombotic risk factors identified Four patients in these studies were reported to have a recurrence of RVT and all had at least one prothrombotic risk factor. Therefore, because of the high yield in finding inherited thrombophilias in patients with RVT, a comprehensive evaluation of these patients is warranted. 22 Acute complications associated with RVT include adrenal hemorrhage and extension of the clot into the IVC. 22 The diagnosis of RVT is usually made by US with Doppler. Sonographic features include enlarged and echogenic kidneys with attenuation or loss of corticomedullary differentiation. The treatment for RVT is controversial as there are no large randomized clinical trials addressing the issue. The most recent recommendations are that for unilateral RVT in the absence of uremia or extension into the IVC, either supportive care with monitoring for extension or anticoagulation with UFH or LMWH for up to 3 months is acceptable. 73 For unilateral RVT with extension into the IVC, anticoagulation with UFH or LMWH is recommended for 3 months. 73 Thrombolytic therapy plus anticoagulation with UFH followed by anticoagulation with UFH or LMWH should be reserved for those with bilateral RVT with renal failure. 73 Supportive of these recommendations, a recent review found that fibrinolytic therapy had no effect on the long-term outcome on the affected kidney in those with unilateral RVT but renal function was restored in two patients with bilateral RVT. 84 Portal venous thrombosis. The true incidence of portal venous thrombosis (PVT) is unknown as it tends to be clinically silent. 18 The risk of PVT is mainly due to sepsis/oomphalitis and UVC use. The incidence of PVT with UVC use has been reported to vary between 1 and 43%. 85 The diagnosis of PVT is made by US and findings of cavernous transformation of the portal vein with subsequent splenomegaly; reversal of portal flow is used to document its severity. 18 Spontaneous resolution of asymptomatic PVT is relatively common. However, if PVT is detected, close observation is important to follow for signs of portal hypertension. This may manifest itself up to 10 years after the neonatal period as splenomegaly without liver disease, reversal of portal vein flow and gastric/esophageal varices. 18 Cerebral sinovenous thrombosis. The incidence of cerebral sinovenous thrombosis (CSVT) in childhood is estimated at 0.67 per children per year with approximately half occurring in neonates. 25,26 Major presenting clinical features of CSVT include seizures, fever and lethargy. 25,26 The most recent review of infants with CSVT found that 90% had predisposing factors identified (Table 2), with 16% of these patients having multiple risk factors. 26 The superior and lateral sinuses are the most frequently involved vessels. 86 Up to 30% of cases have reported venous infarction with subsequent hemorrhage. 87 The diagnosis of CSVT is best achieved through diffusion MRI with venography. 27,28,88 Reported outcomes vary in neonates with CSVT. The Canadian Cohort 89 reported that 8% of neonates with CSVT died, and followup data of these patients found that after a mean duration of

5 Postnatal thromboses years, 77% had no neurological sequelae. 89,90 However, a more recent review found that 6 of 25 (24%) neonates with CSVT died, with death more likely if coma or seizures were present. 26 The reasons for the increased mortality rate in this study were unclear. Recommendations for management (not based on clinical trials) are that either UFH or LMWH be used for the initial period only for neonates without large ischemic infarctions or intracerebral hemorrhage, followed by LMWH for a minimum of 6 weeks but not longer than 3 months. 73 For the remainder, radiological monitoring for 5 to 7 days and supportive care is recommended and anticoagulation should be started only if extension occurs. 73 There is only one report of a neonate with parasagittal hemorrhage due to cortical vein thrombosis who was treated with locally applied urokinase and recovered completely. 91 However, more experience with this modality is needed before this treatment is recommended. Surgery is recommended only for those with hydrocephalus or large intracerebral hematomas with mass effect. 73,86 Clinical risk factors for postnatal thrombosis Acquired risk factors The presence of central venous and arterial catheters is the greatest acquired risk for the development of postnatal TE. 2 4,18,29 Although these types of catheters have improved care for the smallest and most critically ill neonates, they clearly predispose an infant to significant risks for thrombus formation. Reviews evaluating whether effects of catheter materials, position of the catheter tip, catheter design and number of lumens have been performed. These reviews have demonstrated fewer vascular complications associated with high UAC position, fewer thrombotic complications with radiopaque silicone elastomere vs polyvinyl chloride UACs and fewer thrombotic complications associated with end-hole UACs. 58,59,92 95 Despite catheters being the major culprit for the development of postnatal thromboses, other risk factors have been demonstrated. A complete listing of risk factors implicated in the development of postnatal TE is displayed in Table 2. Most patients with reportable thromboses had multiple coexisting risk factors present. 2 4 Genetic risk factors With the increased detection of thromboses in the neonate, more emphasis has been placed on the role of genetic prothrombotic defects. These genetic mutations may result in the complete absence or severe deficiency of an inhibitor of hemostasis, the production of an inhibitor of hemostasis that has inadequate function despite normal levels or an overproduction of a procoagulant protein or cofactor. Homozygous prothrombotic disorders, most likely severe protein C, protein S or antithrombin deficiency, are usually present in newborns with severe clinical manifestations (purpura fulminans) that require specific urgent attention. 73 Major questions exist regarding how to interpret and manage heterozygous congenital prothrombotic disorders and if a neonate with a significant TE event warrants a complete thrombophilia evaluation. 18,73,86,96 Despite a high prevalence of multiple thrombophilic gene mutations in the Caucasian population, 30 a high percentage of neonates do not develop thromboses, even with multiple acquired risk factors. 17 However, many experts believe that screening for the presence of a heterozygous disorder is warranted in a neonate with a significant TE event, regardless of the presence of other risk factors. This argument is based on registry data and case studies that have demonstrated that the majority of symptomatic neonatal thromboses stem either from the association of multiple hemostatic prothrombotic defects or the combination of prothrombotic defects and environmental or clinical conditions. 1 4,6,9,22 24,30 37,62, As a result of this association, the Subcommittee for Perinatal and Pediatric Thrombosis of the Scientific and Standardization Committee of the International Society on Thrombosis and Haemostasis recommend that pediatric patients with thrombosis (regardless of other risk factors) be tested for a full panel of genetic and acquired prothrombotic traits (Tables 3 and 4). 38 The prevalence of the more common disorders is listed in Table 5. The Factor V Leiden mutation decreases the rate of inactivation of activated factor V by its regulator, activated protein C. 99 Individuals that are heterozygous for the prothrombin mutation have the potential to increase circulating concentrations of prothrombin by 15 to 30%, which can significantly increase thrombin generation. 105 Defects in the remethylation of homocysteine into methionine due to a mutation in the methylenetetrahydrofolate reductase (MTHFR) enzyme (MTHFR667 and MTHFR1298 mutations) may lead to hyperhomocysteinemia. 17 This mutation plus the increased Table 3 Laboratory evaluation 1 for prothrombotic disorder a Laboratory test Antiphospholipid antibody panel, anticardiolipin and lupus anticoagulant (IgG, IgM) Fibrinogen Protein C activity Protein S activity Antithrombin (activity assay) Factor V Leiden mutation b Prothrombin G b MTHFR mutation b Homocysteine Lipoprotein-a Collection tube EDTA a The evaluation was performed if other acquired risk factors were present. b DNA-based assay. Adapted with permission. 51

6 472 Postnatal thromboses Table 4 Laboratory evaluation 2 for prothrombotic disorder a Laboratory test Antiphospholipid antibody panel, anticardiolipin and lupus anticoagulant (IgG, IgM) FVIII activity FIX activity FXI activity FXII activity Fibrinogen Protein C activity Protein S activity Antithrombin (activity assay) Plasminogen activity Factor V Leiden mutation b Prothrombin G b MTHFR mutation b Lipoprotein-a Homocysteine Heparin cofactor II Collection tube EDTA Abbreviation: MTHFR, methylenetetrahydrofolate reductase. a The evaluation was performed if other acquired risk factors were not present. b DNA-based assay. Adapted with permission. 51 Table 5 Prevalence of certain prothrombotic risk factors Mutation Prevalence Factor V Leiden mutation 4 6% of Caucasian population Prothrombin 20210G>A mutation 1 2% of Caucasian population (heterozygous) 105,106 MTHFR667 African Americans 107, % (homozygous) % (heterozygous) Caucasians 107, % (homozygous) (heterozygous) MTHFR1298 African Americans 107,108 Increased plasma factor VIII activity 2 4% (homozygous) 27 30% (heterozygous) Caucasians 107, % (homozygous) % (heterozygous) 10% of Caucasian population 109 homocysteine concentration have been suspected to increase premature vascular disease and arterial thrombosis. 97,110 Currently, prospective clinical trials of vitamin therapy seek to determine if hyperhomocysteinemia is a risk factor for atherothrombosis. 109,110 Thus, the MTHFR gene mutation and the risk of pediatric thrombosis remain controversial. 36,111 The protein C system contains protein C, with protein S and factor V as cofactors. Thrombin converts protein C into activated protein C that cleaves and inactivates activated factors V and VIII. Genetic mutations that result in quantitative or qualitative deficiencies of either protein C or protein S may contribute to an increase in venous TE. 35,109 Antithrombin, when bound to heparin sulfate, inhibits thrombin and activated factors XI, IX and X. Genetic mutations that result in deficiencies in antithrombin production may increase the risk for TE. 109 Elevations in the production of factors VIII, IX and XI have been reported and implicated in the development of thromboses. 109,112,113 Although elevations in lipoprotein-a concentrations have been found to increase risk for both venous and arterial thromboses in the German population, 33,34 applying this generalization to all populations remains unclear. The presence of maternal antiphospholipid antibodies, including lupus anticoagulant, anticardiolipin antibody and anti-b2-glycoprotein-i antibodies, may be a cause of postnatal thrombosis. The explanation is that these antibodies are transported across the placenta and may serve as a nidus for thrombosis in the neonate. 16,86,114 Infants born to mothers with a history of antiphospholipid antibodies are recommended to be observed closely for signs and symptoms of thromboses. 16 Evaluation of clinically symptomatic postnatal thromboses The proper evaluation of a neonate with a clinically significant postnatal thrombosis can be somewhat confusing, mainly because there are limited studies that have focused on the necessary evaluation, treatment and follow-up of a neonate with a postnatal TE event. Most recommendations are based on case reports, case series, cohort studies, reviews and expert opinion. 73 In addition, many current practices in neonates are extrapolated from studies carried out in older children and adults. When faced with a neonate that has developed a clinically significant thrombosis in the postnatal period, a detailed family history is recommended in an attempt to identify a potential prothrombotic disorder. Prior pregnancy history in the mother revealing multiple spontaneous abortions, miscarriages or premature births are strong clues for an inherited thrombophilia. In addition, a detailed pregnancy, delivery and immediate postnatal history are important to identify potential risk factors (Table 2). Although an extensive laboratory search for a prothrombotic disorder is controversial, 17,18,86,96 it is our opinion that the weight of evidence 1 4,6,9,22 24,30 37,62, supports a step-wise thrombophilia investigation in the presence of a clinically significant postnatal thromboembolic event, regardless of the

7 Postnatal thromboses 473 number of other predisposing risk factors (Tables 3 and 4). The protein-based assays (Tables 3 and 4) are affected by the thrombotic event and must be repeated at 3 to 6 months of life before a definitive diagnosis can be made. 6,88 At our institution, we perform laboratory evaluation 1 (Table 3) if other significant risk factors are present (Table 2). This evaluation is focused on the more common prothrombotic risk factors. In the absence of any risk factors, then laboratory evaluation 2 is performed (Table 4). The clinician can choose to add or delete tests based on the presentation, family history or time frame of the event. The evaluation is performed within several days of detection of the thrombosis and necessary testing is repeated at 3 to 6 months of life. In addition, if anticoagulation therapy is being used, then the protein-based assays are recommended to be obtained 14 to 30 days after discontinuing the agent. DNA-based assays are neither influenced by the acute thrombotic event nor the treatment and are reliable if obtained during the acute event. 6 Lipoprotein-a concentrations tend to increase during the first year of life and are recommended to be repeated at 8 to 12 months of life following the acute event if the original values are low, especially in Caucasian individuals. 6,33,100 A concern to performing this evaluation is the large amount of blood believed to be required. This is especially important in the premature or anemic infant where excessive blood loss is not a tolerable option. Because of this, we feel that the laboratory evaluation should take place at an experienced tertiary care center that has either self-laboratory support or a reliable referral center. This allows the most comprehensible evaluation be performed with a minimal blood requirement. Management of thrombosis The most concerning complication associated with antithrombotic therapy in the NICU is intracranial hemorrhage. Recommendations and dosing regimens for antithrombotic therapy in neonates are based on case series, cohort studies and expert opinion. 73 Before initiating antithrombotic therapy, the clinician must consider the potential for serious complications and assume that the treatment s benefits will outweigh its risks, especially in the premature infant. Treatment must occur in a tertiary care center that has proper laboratory, blood bank, hematological subspecialty and pediatric surgical support. The clinician can also use the service, NO CLOTS, to receive upto-date management guidance. Absolute contraindications for antithrombotic therapy include central nervous system surgery or ischemia (including birth asphyxia within 10 days), active bleeding, invasive procedures within 3 days and seizures within 48 h. Contraindications include a platelet count < per microliter, fibrinogen concentration <100 mg per 100 ml, INR >2, a severe coagulation deficiency and known allergy to an antithrombotic agent. 7,10,18,73,88 Anticoagulant therapy When faced with the decision for neonatal anticoagulation, the choice is either UFH or LMWH. UFH is an attractive anticoagulant for neonates due to its short half-life allowing for rapid dose adjustments and discontinuation of therapy. However, the neonatal response to heparin is affected by their low levels of antithrombin (especially in the sick, premature infant) and increased rate of heparin clearance. 39 These issues can lead to neonates being somewhat resistant to the anticoagulant effects of UFH and thus higher doses are frequently required to achieve therapeutic levels. 39,88 UFH requires 24 h i.v. access and concerns for side effects such as heparin-induced thrombocytopenia (HIT) exist. LMWH therapy is ideal for neonatal use in that it has a reduced risk of hemorrhage, does not require venous access and requires reduced monitoring requirements. 7 Other side effects such as HIT are reduced in children and adults, and theoretically reduced in neonates. 115,116 In addition, the placement of a subcutaneous catheter can dramatically reduce the number of subcutaneous injections. However, LMWH has a longer half-life and 12 to 24 h is recommended after the last dose before any invasive procedures can occur. Owing to the above risks and benefits for each agent, we recommended UFH be used only if short-term anticoagulation is desired, especially if surgery is indicated within the next 72 h. If long-term anticoagulation is desired, then LMWH is a more practical choice in the NICU. Unfractionated heparin. The use of UFH in neonates is recommended for clinically significant or large thromboses that are not limb or life threatening with the goal of preventing clot expansion or embolism (see prior sections). Traditional dosing plus more recent dosing guidelines based on gestational age are displayed in Table 6. Regardless of dosing used, the goal of treatment is to maintain an anti-factor Xa (anti-fxa) level of 0.3 to 0.7 U ml 1 (corresponds to APTT of 60 to 85 s). 16,88,117,118 An anti-fxa level should be checked 4 h after the loading dose and then 4 h after each change made in the infusion rate. Complete blood count and APTT testing should be done before starting UFH therapy and repeated daily once levels are maintained. Therapy is usually continued for 5 to 30 days, 73 but data to support this recommendation are lacking. The most significant complication of UFH is bleeding. The German registry of neonatal thrombosis reported rates of major hemorrhage of 2% for heparin anticoagulation. 4 If this occurs, simple cessation of the infusion will treat this due to the short halflife of UFH. However, if bleeding continues, then a full coagulation assessment should be performed and hemostatic deficiencies should be replaced as indicated. If the anti-fxa activity is >0.8 U ml 1 in the face of active bleeding, then protamine may be considered. Suggested dosing for protamine is 1 mg per 100 U heparin received if the last heparin dose was administered less than 30 min

8 474 Postnatal thromboses Table 6 Recommended therapy and dosing of neonatal antithrombotic agents Clinical indication Medication Traditional dosing Current recommended dosing Asymptomatic or symptomatic thrombus but non-limb threatening Asymptomatic or symptomatic thrombus but non-limb threatening UFH LMWH Traditional dosing: 75 U kg 1 i.v. over 10 min, then 28 U kg 1 per h Traditional dosing: 1.5 mg kg 1 SQ per 12 h <28-week gestation: 25 U kg 1 i.v. over 10 min, then 15 U kg 1 per h 28- to 37-week gestation: 50 U kg 1 over 10 min, then 15 U kg 1 per h >37-week gestation: 100 U kg 1 over 10 min, then 28 U kg 1 per h Term neonates 1.7 mg kg 1 SQ per 12 h Preterm neonates 2.0 mg kg 1 SQ per 12 h Limb/life-threatening thrombus rt-tpa NA 0.06 mg kg 1 per h UFH at 10 U kg 1 per h Abbreviation: MTHFR, methylenetetrahydrofolate reductase; NA, not applicable; SQ, subcutaneous; rt-tpa, recombinant tissue plasminogen activator; UFH, unfractionated heparin. Adapted with permission. 16 previously; 0.5 to 0.75 mg per 100 U heparin received if the last heparin dose was administered 30 to 60 min previously; to 0.5 mg per 100 U heparin received if the last heparin dose was administered 60 to 120 min previously and 0.25 to mg per 100 U heparin received if the last dose was administered >120 min previously. 73,119 Protamine dosing should be conservative as excess protamine is also an anticoagulant. Although a significant concern in adults, there have only been a few cases reported of HIT in newborns. 120,121 A drop in the platelet count by 50% or persistent platelet counts <70 to per mm 3 occurring 5 to 10 days after the first exposure to heparin should alert the clinician for this diagnosis. Lepirudin can be considered in neonates with HIT, but its use has not been adequately studied in neonates. 7 Low-molecular-weight heparin Although different LMWH preparations exist (enoxaparin, dalteparin, reviparin), all have a specific activity in vitro against FXa. Since the 1990s, LMWH, specifically enoxaparin, has become the anticoagulant of choice in NICUs. 122,123 The administration of LMWH can be achieved either through subcutaneous injection or through the use of an indwelling subcutaneous catheter (Insuflon; Unomedical, Birkerod, Denmark). 10,88,122 The use of this catheter reduces the number of needle sticks from 14 per week to 1 per week. 124 A recent review found that over 50% of patients who had subcutaneous catheters placed experienced minor adverse events that included induration, leakage and bruising. Only one patient, a 24-week infant, required antibiotic treatment secondary to an infected hematoma. 125 LMWH therapy has been effective in the NICU, and centers have reported either partial or complete resolution of TE events between 59 and 100%. 125,126 The traditional initial treatment dose of enoxaparin is displayed in Table 6. A recent review evaluating enoxaparin use in 240 neonates found that the mean maintenance dose of enoxaparin ranged from 1.48 to 2.27 mg kg 1 per 12 h for all infants, but was higher for preterm neonates at 1.9 to 2.27 mg kg 1 per 12 h. 125 Therefore, newer recommendations are displayed in Table 6. The goal of treatment is to maintain an anti-fxa concentration of 0.5 to 1.0 U ml 1. Levels should be obtained 4 h after the second dose and then weekly (obtained 4 h after a specified dose). Guidelines for adjusting LMWH therapy are published in other sources. 10,73,88 Anti-FXa levels can be obtained through heparinized arterial or venous catheters using a port with no heparin running through it, if 4.0 ml of blood is cleared from the line before obtaining the sample. 127 Thrombolytic agents Thrombolysis should only be considered for limb- or organthreatening postnatal thrombosis and acute atrial clots. 7,73,88,128 Recombinant tissue plasminogen activator (rt-tpa) is the most widely used agent and recommended for NICU use. 79 rt-tpa has several advantages for newborns, including short half-life, minimal antigenicity, direct activation of plasminogen, lack of inhibition by a 2 antiplasmin and localization of fibrinolytic activity. 73,129 However, no randomized trials comparing its safety and efficacy to other agents have been performed in neonates. Over the past 8 years, five case series were published evaluating different protocols for neonatal rt-tpa therapy. 10,79 Although different protocols were used, complete lysis of clots was noted in 31 of 50 patients with partial lysis in another 11 patients. 10,79 Included in this series are the Pediatric Coagulation Consortium reviewed results from five neonates treated with low-dose rt-tpa. Treatment was effective in all neonates treated and only one preterm neonate had therapy discontinued due to the development of a subdural hematoma. 79 On the basis of these limited patients, dosing recommendations are displayed in Table 6. 79,88 Dose escalation up to 0.24 mg kg 1 per h can be considered but done slowly with continuing monitoring of the patient. Owing to the fact that thrombolysis does not inhibit clot propagation or directly affect hypercoagulability, simultaneous infusion of UFH at 10 U kg 1 per h is recommended. 79,88

9 Postnatal thromboses 475 Table 7 Laboratory monitoring during thrombolytic therapy 7,10,79,88,130 Laboratory value a Levels desired Fibrinogen Minimum of 100 mg/100 ml 117 Supplement with cryoprecipitate Platelet count Minimum of Plasminogen/D-dimers Adequate to achieve thrombolysis Adapted with permission. 51 a Levels should be obtained before the initiation of treatment, 4 6 h after staring treatment and every h during treatment. Monitoring of thrombolysis is critical and requires imaging studies of the clot every 4 to 24 h (can do every 12 to 24 h after initial 24 h of therapy). 73,79,88 In addition, cranial US should be performed before initiation of treatment and then daily during thrombolytic therapy. 7 Once clot lysis is achieved, rt-tpa therapy should be discontinued. Coagulation values should be obtained before the initiation of treatment and followed every 12 to 24 h during treatment to assure that no significant coagulopathy is occurring. Further laboratory management guidelines are listed in Table 7. Surgery The goal for surgical intervention is to maximize the chances of limb salvage and normal limb growth. The use of microsurgical techniques combined with thrombolytic regimens has the potential to rapidly restore blood flow, avoiding tissue loss, without major bleeding complications, especially in patients with peripheral arterial occlusion. 131 A recent review of one center s experience of 11 patients with arterial vascular access-associated thrombosis described 5 patients that required arteriotomy, embolectomy and subsequent microvascular reconstruction. 131 When faced with a significant arterial thrombosis, especially one from a PAL, surgery may be considered. New anticoagulants New anticoagulant agents have been released for use in adults, but little experience in children, especially neonates, has been reported. Recombinant hirudin (lepirudin) is a direct thrombin inhibitor and can be used in neonates with HIT. 7 The dose for newborns with HIT is 0.03 to 0.05 mg kg 1 per h. Dosing is monitored with the Ecarin clotting time. The direct oral thrombin inhibitor, Melagatran, has potential in the NICU but remains to be tested. 132 In addition, new catheter materials in the clinical and preclinical stage of development such as those coated with antithrombin heparin covalent complex may prove useful. 18 Future The lack of randomized clinical trials evaluating the management of postnatal thromboses coupled with the potential risks of treatment creates a difficult environment with how to properly treat neonates with clinically significant postnatal thromboses. In addition, applying retrospective data and case reports to every situation may not be reflective of the population that is being cared for, especially considering that successful treatments are more likely to be reported by authors than are complications of therapy. Databases established to report site and population-specific trends should become a mainstay of care in the future. With the potential for newer and safer catheters and antithrombotic therapy, it is hoped that complications can be minimized and outcomes can be improved. Performing focused prothrombotic evaluations will not only minimize unnecessary testing, but will allow for the identification of potential genetic disorders to help avoid future thrombotic catastrophes. Acknowledgments This work was supported by a grant from the American Heart Association, Florida Affiliate. References 1 Kenet G, Nowak-Gottl U. Fetal and neonatal thrombophilia. Obstet Gynecol Clin North Am 2006; 33: van Ommen CH, Heijboer H, Buller HR, Hirasing RA, Heijmans HS, Peters M. Venous thromboembolism in childhood: a prospective two-year registry in the Netherlands. J Pediatr 2001; 139: Schmidt B, Andrew M. Neonatal thrombosis: report of a prospective Canadian and international registry. Pediatrics 1995; 96: Nowak-Gottl U, von Kries R, Gobel U. Neonatal symptomatic thromboembolism in Germany: two year survey. Arch Dis Child Fetal Neonatal Ed 1997; 76: F163 F Kurnik K, Kosch A, Strater R, Schobess R, Heller C, Nowak-Gottl U. Recurrent thromboembolism in infants and children suffering from symptomatic neonatal arterial stroke: a prospective follow-up study. Stroke 2003; 34: Nowak-Gottl U, Duering C, Kempf-Bielack B, Strater R. Thromboembolic diseases in neonates and children. Pathophysiol Haemost Thromb 2003; 33: Thornburg C, Pipe S. Neonatal thromboembolic emergencies. Semin Fetal Neonatal Med 2006; 11: Andrew M, Paes B, Milner R, Johnston M, Mitchell L, Tollefsen DM et al. Development of the human coagulation system in the healthy premature infant. Blood 1988; 72: Manco-Johnson MJ. Development of hemostasis in the fetus. Thromb Res 2005; 115(Suppl 1): Ramasethu J. Management of vascular thrombosis and spasm in the newborn. Neoreviews 2005; 6: e298 e Saxonhouse M, Sola-Visner M. Current issues in the pathogenesis, diagnosis, and treatment of neonatal thrombocytopenia. In: Ohls R, Yoder M (eds). Hematology, Immunology and Infectious Disease: Neonatology Questions and Controversies, vol 1 Saunders Elsevier: Philadelphia, 2008, pp Kettner SC, Pollak A, Zimpfer M, Seybold T, Prusa AR, Herkner K et al. Heparinasemodified thrombelastography in term and preterm neonates. Anesth Analg 2004; 98: , table of contents. 13 Andrew M, Paes B, Johnston M. Development of the hemostatic system in the neonate and young infant. Am J Pediatr Hematol Oncol 1990; 12: Manco-Johnson MJ. Etiopathogenesis of pediatric thrombosis. Hematology 2005; 10(Suppl 1):

10 476 Postnatal thromboses 15 Journeycake JM, Manco-Johnson MJ. Thrombosis during infancy and childhood: what we know and what we do not know. Hematol Oncol Clin North Am 2004; 18: , viii ix. 16 Manco-Johnson MJ. Controversies in neonatal thrombotic disorders. In: Ohls R, MC Y (eds). Hematology, Immunology, and Infectious Disease: Neonatology Questions and Controversies, vol 1 Saunders Elsevier: Philadelphia, 2008, pp Turebylu R, Salis R, Erbe R, Martin D, Lakshminrusimha S, Ryan RM. Genetic prothrombotic mutations are common in neonates but are not associated with umbilical catheter-associated thrombosis. J Perinatol 2007; 27: Greenway A, Massicotte MP, Monagle P. Neonatal thrombosis and its treatment. Blood Rev 2004; 18: Vailas GN, Brouillette RT, Scott JP, Shkolnik A, Conway J, Wiringa K. Neonatal aortic thrombosis: recent experience. J Pediatr 1986; 109: Tanke RB, van Megen R, Daniels O. Thrombus detection on central venous catheters in the neonatal intensive care unit. Angiology 1994; 45: Schmidt B, Zipursky A. Thrombotic disease in newborn infants. Clin Perinatol 1984; 11: Lau KK, Stoffman JM, Williams S, McCusker P, Brandao L, Patel S et al. Neonatal renal vein thrombosis: review of the English-language literature between 1992 and Pediatrics 2007; 120: e1278 e Kosch A, Kuwertz-Broking E, Heller C, Kurnik K, Schobess R, Nowak-Gottl U. Renal venous thrombosis in neonates: prothrombotic risk factors and long-term follow-up. Blood 2004; 104: Marks SD, Massicotte MP, Steele BT, Matsell DG, Filler G, Shah PS et al. Neonatal renal venous thrombosis: clinical outcomes and prevalence of prothrombotic disorders. J Pediatr 2005; 146: deveber G, Andrew M, Adams C, Bjornson B, Booth F, Buckley DJ et al. Cerebral sinovenous thrombosis in children. N Engl J Med 2001; 345: Wasay M, Dai AI, Ansari M, Shaikh Z, Roach ES. Cerebral venous sinus thrombosis in children: a multicenter cohort from the United States. J Child Neurol 2008; 23: Shroff M, deveber G. Sinovenous thrombosis in children. Neuroimaging Clin N Am 2003; 13: de Veber G, Adams M, Andrew M. Cerebral thromboembolism in neonates: clinical and radiographic features. Blood 1998; 92: Beardsley DS. Venous thromboembolism in the neonatal period. Semin Perinatol 2007; 31: Rosendaal FR. Venous thrombosis: the role of genes, environment, and behavior. Hematology (Am Soc Hematol Educ Program) 2005; Seligsohn U, Zivelin A. Thrombophilia as a multigenic disorder. Thromb Haemost 1997; 78: Nowak-Gottl U, Junker R, Kreuz W, von Eckardstein A, Kosch A, Nohe N et al. Risk of recurrent venous thrombosis in children with combined prothrombotic risk factors. Blood 2001; 97: Nowak-Gottl U, Strater R, Heinecke A, Junker R, Koch HG, Schuierer G et al. Lipoprotein (a) and genetic polymorphisms of clotting factor V, prothrombin, and methylenetetrahydrofolate reductase are risk factors of spontaneous ischemic stroke in childhood. Blood 1999; 94: Nowak-Gottl U, Junker R, Hartmeier M, Koch HG, Munchow N, Assmann G et al. Increased lipoprotein(a) is an important risk factor for venous thromboembolism in childhood. Circulation 1999; 100: Bucciarelli P, Rosendaal FR, Tripodi A, Mannucci PM, De Stefano V, Palareti G et al. Risk of venous thromboembolism and clinical manifestations in carriers of antithrombin, protein C, protein S deficiency, or activated protein C resistance: a multicenter collaborative family study. Arterioscler Thromb Vasc Biol 1999; 19: Brenner B. Thrombophilia and adverse pregnancy outcome. Obstet Gynecol Clin North Am 2006; 33: , ix. 37 Boffa MC, Lachassinne E. Infant perinatal thrombosis and antiphospholipid antibodies: a review. Lupus 2007; 16: Manco-Johnson MJ, Grabowski EF, Hellgreen M, Kemahli AS, Massicotte MP, Muntean W et al. Laboratory testing for thrombophilia in pediatric patients. On behalf of the Subcommittee for Perinatal and Pediatric Thrombosis of the Scientific and Standardization Committee of the International Society of Thrombosis and Haemostasis (ISTH). Thromb Haemost 2002; 88: Male C, Johnston M, Sparling C, Brooker L, Andrew M, Massicotte P. The influence of developmental haemostasis on the laboratory diagnosis and management of haemostatic disorders during infancy and childhood. Clin Lab Med 1999; 19: Wu YW, Lynch JK, Nelson KB. Perinatal arterial stroke: understanding mechanisms and outcomes. Semin Neurol 2005; 25: Raju TN, Nelson KB, Ferriero D, Lynch JK. Ischemic perinatal stroke: summary of a workshop sponsored by the National Institute of Child Health and Human Development and the National Institute of Neurological Disorders and Stroke. Pediatrics 2007; 120: Martinelli I, Taioli E, Cetin I, Marinoni A, Gerosa S, Villa MV et al. Mutations in coagulation factors in women with unexplained late fetal loss. N Engl J Med 2000; 343: Manco-Johnson MJ, Abshire TC, Jacobson LJ, Marlar RA. Severe neonatal protein C deficiency: prevalence and thrombotic risk. J Pediatr 1991; 119: Lynch JK, Nelson KB. Epidemiology of perinatal stroke. Curr Opin Pediatr 2001; 13: Leistra-Leistra MJ, Timmer A, van Spronsen FJ, Geven WB, van der Meer J, Erwich JJ. Fetal thrombotic vasculopathy in the placenta: a thrombophilic connection between pregnancy complications and neonatal thrombosis? Placenta 2004; 25(Suppl A): S102 S Lee J, Croen LA, Backstrand KH, Yoshida CK, Henning LH, Lindan C et al. Maternal and infant characteristics associated with perinatal arterial stroke in the infant. JAMA 2005; 293: Laugesaar R, Kolk A, Tomberg T, Metsvaht T, Lintrop M, Varendi H et al. Acutely and retrospectively diagnosed perinatal stroke: a population-based study. Stroke 2007; 38: Grandone E, Margaglione M, Colaizzo D, D Andrea G, Cappucci G, Brancaccio V et al. Genetic susceptibility to pregnancy-related venous thromboembolism: roles of factor V Leiden, prothrombin G20210A, and methylenetetrahydrofolate reductase C677T mutations. Am J Obstet Gynecol 1998; 179: den Heijer M, Koster T, Blom HJ, Bos GM, Briet E, Reitsma PH et al. Hyperhomocysteinemia as a risk factor for deep-vein thrombosis. N Engl J Med 1996; 334: de Veber G. Canadian pediatric iscaemic stroke registry. Pediatr Child Health 2000b; A Saxonhouse M, Manco-Johnson M. The evaluation and management of neonatal coagulation disorders. Semin Perinatol 2009; 33: Ramasethu J. Complications of vascular catheters in the neonatal intensive care unit. Clin Perinatol 2008; 35: , x. 53 Fletcher MA, Brown DR, Landers S, Seguin J. Umbilical arterial catheter use: report of an audit conducted by the Study Group for Complications of Perinatal Care. Am J Perinatol 1994; 11: Seibert JJ, Northington FJ, Miers JF, Taylor BJ. Aortic thrombosis after umbilical artery catheterization in neonates: prevalence of complications on long-term follow-up. AJR Am J Roentgenol 1991; 156: Nouri S, Mahdhaoui N, Beizig S, Salem N, Methlouthi J, Ben Dhafer S et al. [Major neonatal aortic thrombosis: a case report]. Arch Pediatr 2007; 14: Joshi VV, Draper DA, Bates III RD. Neonatal necrotizing enterocolitis. Occurrence secondary to thrombosis of abdominal aorta following umbilical arterial catheterization. Arch Pathol 1975; 99: Greenberg R, Waldman D, Brooks C, Ouriel K, Pegoli W, Ryan R et al. Endovascular treatment of renal artery thrombosis caused by umbilical artery catheterization. J Vasc Surg 1998; 28: Barrington KJ. Umbilical artery catheters in the newborn: effects of position of the catheter tip. Cochrane Database Syst Rev 2000 CD