Fax: (416) 813-5979
A prospective cohort study determining the prevalence of thrombotic events in children with acute lymphoblastic leukemia and a central venous line who are treated with L-asparaginase†
Results of the Prophylactic Antithrombin Replacement in Kids with Acute Lymphoblastic Leukemia Treated with Asparaginase (PARKAA) Study
Article first published online: 3 JAN 2003
Copyright © 2003 American Cancer Society
Volume 97, Issue 2, pages 508–516, 15 January 2003
How to Cite
Mitchell, L. G. and and the PARKAA Group (2003), A prospective cohort study determining the prevalence of thrombotic events in children with acute lymphoblastic leukemia and a central venous line who are treated with L-asparaginase. Cancer, 97: 508–516. doi: 10.1002/cncr.11042
The authors are all deeply saddened by the loss of their colleague and friend, Dr. Maureen Andrew, who passed away suddenly on August 28, 2001. Dr. Andrew was instrumental in the work reported here, and the authors respectfully dedicate this article to her memory.
- Issue published online: 3 JAN 2003
- Article first published online: 3 JAN 2003
- Manuscript Accepted: 15 AUG 2002
- Manuscript Revised: 12 AUG 2002
- Manuscript Received: 22 FEB 2002
- Canadian Institutes of Health Research. Grant Number: PA 14283
- Bayer Inc.
- acute lymphoblastic leukemia;
Thrombotic events (TEs) are serious secondary complications in children with acute lymphoblastic leukemia (ALL) who receive L-asparaginase (ASP) therapy; however, the prevalence of TEs has not been established. The primary objective of the Prophylactic Antithrombin Replacement in Kids with Acute Lymphoblastic Leukemia Treated with Asparaginase (PARKAA) Study was to determine the prevalence of TEs. The secondary objective was to detect any association of TEs with the presence of congenital or acquired prothrombotic disorders.
Children with ALL were screened for TEs at the end of ASP treatment using bilateral venograms, ultrasound, magnetic resonance imaging, and echocardiography. Symptomatic TEs were confirmed by appropriate radiographic tests. All tests were read by a blinded central adjudication committee.
Twenty-two of 60 children had TEs, a prevalence of 36.7% (95% confidence interval, 24.4–48.8%). TEs were located in the sinovenous system of the brain in 1 patient, the right atrium in 3 patients, and the upper central venous system in 19 patients. TEs detected by venography resulted in 1) 25–100% occlusion, with 1 in 3 patients showing occlusion of > 75% of the greatest vessel dimension, and 2) the presence of collaterals in 60% of patients, with 40% categorized as major. No children with TEs were positive for factor V Leiden or prothrombin gene 20201A, and four of eight children with antiphospholipid antibodies had a TE.
The prevalence of TEs is exceedingly high in this population, and it is likely that the extent of occlusion is likely clinically significant. No trend was seen toward an association between TEs and the presence of congenital prothrombotic disorders. A trend was seen toward an association between TEs and antiphospholipid antibodies. Carefully designed clinical trials of primary prophylaxis for the prevention of TEs are required in this patient population. Cancer 2003;97:508–16. © 2003 American Cancer Society.
Acute lymphoblastic leukemia (ALL) occurs in approximately 2500 North American children per year. The treatment of patients with childhood ALL has advanced to the extent that 5-year event free survival rates are approximately 80%.1 Because survival rates are excellent, the morbidities secondary to treatment for ALL have assumed increasing importance. Thromboembolic events (TEs) are among the more frequent and serious complications of ALL and its treatment. The timing of TEs in children with ALL is very consistent in the literature, occurring either during or immediately after chemotherapy with L-asparaginase (ASP).2–14
ASP is an effective chemotherapeutic agent that catalyzes the hydrolysis of L-asparagine. Human lymphoblasts rely on exogenous asparagine, and ASP induces a relative asparagine deficiency, resulting in cell death.15 ASP causes cessation of protein biosynthesis in tissues, such as the liver, where exogenous asparagine is required. Treatment with ASP decreases plasma concentrations of most coagulation proteins, with the main thrombin inhibitor, antithrombin, reduced to the greatest extent.7, 8, 14, 16 The available information suggests that the acquired antithrombin deficiency in children with ALL increases the risk of TEs. The precise prevalence of TEs remains unknown, because no study has studied prospectively the prevalence of TEs using sensitive, objective testing.4, 7–13 Determination of the true prevalence of TEs is important to ascertain the relevance of prophylactic clinical trials focused on reducing TEs in this patient population.
Although children with ALL who are treated with ASP have an increased risk for TEs, not all patients develop these complications. An association of markers for TEs may be beneficial and may facilitate strategies focused on children at greatest risk. There are at least three potential independent risk factors for TEs in pediatric patients: the presence of factor V Leiden, prothrombin gene 20210A, and antiphospholipid antibodies (APLAs).17–24 Factor V Leiden is a mutation in the factor V (FV), which renders FVa resistant to inactivation by activated protein C (APC).25 FV Leiden is present in 5% of the Caucasian population and in approximately 40% of adults with idiopathic TEs.26, 27 Prothrombin gene 20201A is also a congenital prothrombotic disorder secondary to a mutation in the prothrombin gene that results in increased plasma concentrations of prothrombin.28 Prothrombin gene 20201A is present in 2% of the normal Caucasian population and in 6% of adult patients with idiopathic deep vein thrombosis.28–30 APLAs are acquired prothrombotic risk factors that are related significantly to the presence of TEs in children with systemic lupus erythematosus (SLE) and the APLA syndrome.22, 31 The predictive value of these prothrombotic markers in children with ALL who are receiving ASP is unknown.
The current report presents the primary results of the Prophylactic Antithrombin Replacement in Kids with ALL Treated with Asparaginase (PARKAA) Study, which was undertaken to determine the prevalence of TEs in children with ALL who were receiving treatment with ASP. The secondary outcomes also are reported herein to determine whether there was a trend toward an association between TEs and the presence of the prothrombotic disorders, FV Leiden, prothrombin gene 20210A, or the presence of APLAs. The PARKAA Study was an extended Phase II study, and the third outcome investigating whether there was a trend toward the safety and efficacy of antithrombin replacement will be reported in another article.
MATERIALS AND METHODS
The PARKAA Study was an open, randomized, controlled, extended Phase II trial in pediatric patients with ALL. Patients were allocated in a ratio of 2:1 (for every 2 patients who received no antithrombin supplementation, 1 patient received antithrombin supplementation). The primary objective was to determine the prevalence of thrombosis in the patients that received no antithrombin supplementation. The second objective was to assess the relation of TEs to the presence of the prothrombotic disorders, FV Leiden, prothrombin gene 20210A, or the presence of APLAs. The third objective of the PARKAA Study was to acquire some preliminary data on the efficacy and safety of prophylaxis with antithrombin, as this would be the first use of the product in this population in a clinical study. The study was not powered to prove the efficacy or safety of antithrombin supplementation but, rather, to look for trends. The results of the antithrombin treatment arm will be reported elsewhere. The sample size of the PARKAA Study was based on the primary outcome of prevalence of TE in the nontreated arm; therefore, twice the numbers of patients were randomized to receive no treatment. Only data on children who were randomized to receive no treatment are described in this report. The total sample size of 60 patients was calculated a priori to achieve an estimate of the prevalence of TE with acceptable confidence intervals (95% confidence interval [95%CI]). To ensure that the study findings were generalizable to the patient population, patients from centers that used either Pediatric Oncology Group (POG) protocols or Children's Cancer Group (CCG) protocols were eligible for the study. All centers used exclusively either the POG protocols or the CCG protocols. To minimize bias due to imbalance in the treatment arms, patients were block randomized on center. Patients were stratified on risk of recurrent ALL (high or standard) and randomized in blocks of three by hospital. For the primary outcome, patients had radiographic tests performed after the completion of induction chemotherapy, which included treatment with ASP. The dose of ASP for the POG protocol was 6000 units/m2 on Days 2, 5, 8, 12, 15, and 19 of induction therapy; for the CCG protocol, the ASP dose was 6000 units/m2 on Mondays, Wednesdays, and Fridays for 3 weeks during induction therapy. Patients who developed allergic reactions or had severe pancreatitis associated with ASP were removed from study, as either determined by the attending physician or if ASP was discontinued.
The study population consisted of unselected pediatric patients who were newly diagnosed with ALL. All patients were enrolled prior to the initiation of induction chemotherapy. The inclusion criteria were: age > 6 months and < 18 years, newly diagnosed patients with ALL at the beginning of induction chemotherapy (which included ASP), a functioning central venous line (CVL) placed within 2 weeks of initiating induction chemotherapy, and obtaining informed consent. The exclusion criteria were: previous treatment with ASP, a known hypersensitivity to any of the ingredients in antithrombin concentrate, medical conditions that may have interfered with participation or assessment of the study drug, received other investigational drugs within 30 days of enrollment, or required treatment with therapeutic anticoagulation. The study protocol was reviewed by the Institutional Review Boards of all participating pediatric centers, and informed consent was obtained from all guardians (and from children of an appropriate age).
Assays for Prothrombotic Risk Factors
Blood samples were obtained on Study Day 1 for measurement of FV Leiden, prothrombin gene 20210, and APLA with follow-up samples obtained at the end of the study for a second measurement of APLA on Day 27. Plasmatic antithrombin levels were measured on the days of ASP infusion (Study Days 1, 8, 15, and 22), during the week between ASP infusion (Study Days 4, 11, and 22) and once during follow-up (Study Days 28–35). All samples were assayed at the Pediatric Hemostasis Research Laboratory at the Hospital for Sick Children in Toronto.
Detection of APLAs consisted of measurement of anticardiolipin antibodies and lupus anticoagulants following current recommendations.32 The FV Leiden gene mutation and the prothrombin 20210A gene mutation were analyzed following standard techniques.28, 33 Antithrombin activity levels were measured using a heparin cofactor assay, Chromogenix (DiaPharma, West Chester, OH).
The primary outcome was a clinically symptomatic or asymptomatic TE in any location. TEs were categorized as either not clinically significant: a fibrin sheath or loss of CVL patency that was restored with urokinase [UK] therapy or clinically significant: all other TEs. The degree of vessel occlusion was reported when possible.
Symptomatic TEs were detected by close monitoring of patients during the study period and were confirmed using appropriate objective radiographic tests. No definitions were stipulated for clinical presentation, which was left to the judgment of the attending physician. However, loss of CVL patency was defined as follows: the inability to either infuse or withdraw a sample that could not be corrected by the use of local UK or removal of the CVL due to loss of patency.
Asymptomatic TEs were detected by radiographic testing after the completion of the induction phase of chemotherapy. During follow-up (from Day 28 to 3 months), all patients were screened with radiographic tests, consisting of bilateral venography or magnetic resonance imaging (MRI) of the upper body, ultrasound of the upper body, an echocardiogram, and MRI of the head. Protocols for performance and interpretation of all radiographic tests were defined a priori. All radiographic tests were assessed for presence or absence of TEs by a central adjudication committee, which consisted of physicians who had appropriate expertise, who were not involved in the patient's care, and who were blinded to treatment groups.
The primary analysis was performed on a per protocol basis. The reasons for exclusion from the per protocol population were 1) no exit venogram and 2) premature withdrawal from the study unless due to a TE. For the primary outcome, the point estimate and corresponding 95%CI of TE prevalence was calculated. FV Leiden, prothrombin gene 20210A, and APLA were analyzed as dichotomous data and were compared with the presence of a TE using 2 × 2 contingency tables.
The study was conducted at 10 pediatric tertiary care centers in Canada and the United States. Of the 72 patients who were enrolled in the no antithrombin treatment arm, 67 patients (93.1%) completed the study. Three patients discontinued the study prematurely because of withdrawal of consent, one patient had an adverse event, and eight patients were excluded due to the absence of an exit venogram. Therefore, analysis was performed on a total of 60 of 72 enrolled patients (83%).
Prevalence and Location of Thromboembolic Events
Twenty-two children had confirmed TEs, for a prevalence rate of 36.7% (95%CI, 24.4–48.8%). Locations of the TEs were the sinovenous system of the brain in 1 patient, the right atrium in 3 patients, and the upper central venous system in 19 patients (Figs. 1, 2). One patient had a TE in both the upper system and the right atrium. The majority of TEs, as detected by venography, showed 25–100% vessel occlusion. Collaterals were present in 60% of patients with positive venograms and were categorized as major in 40% (Fig. 2). Three patients (5%; 95%CI, 1–14%) presented with clinically symptomatic TEs, which included swelling and pain in a limb in one patient, extraocular movement abnormalities and headache in one patient, and a blocked line that was unable to be cleared with UK in one patient.
Demographic Data Comparing Patients With and Without TEs
A summary of the demographic data on the 22 patients who developed TEs and the 38 patients who did not have TEs is provided in Table 1. The groups were similar with regard to age, weight, gender, race, and standard/high-risk ALL.
|Characteristic||Thromboembolic event (n = 22)||No thromboembolitic event (n = 38)|
|Female||11 (50%)||21 (55%)|
|Caucasian||17 (77%)||27 (71%)|
|Black||2 (9%)||4 (11%)|
|Other||3 (14%)||7 (18%)|
|High-risk leukemia||9 (41%)||14 (37%)|
Clinical and Laboratory Data in Patients With and Without TEs
There were nonsignificant increased rates of infection and UK use among patients who had TEs compared with patients who did not have TEs (Table 2). The lowest antithrombin level was measured on Day 13 (median; range, Days 2–28) in the clot group and on Day 12 (median; range, Days 2–30) in the no clot group. There was no significant difference in the mean lowest plasma concentration of antithrombin in patients who developed TEs (0.65 ± 0.18 units/mL) and patients who did not develop TEs (0.62 ± 0.16 units/mL).
|Characteristic||Thromboembolic event (n = 22)||No thromboembolitic event (n = 38)|
|%||95% CI||%||95% CI|
Presence of FV Leiden, Prothrombin Gene 20210A, and APLA: Association with Patients at Risk for TEs
The association between TEs and the presence of FV Leiden, prothrombin gene 20210A, and APLA is summarized in Table 3. All patients were assessed for FV Leiden and APLA, because these tests can be performed on plasma samples. However, the prothrombin gene 20210A was tested in only 51 of 60 patients, because no cell samples were received from some patients by the central laboratory. The overall prevalence of the abnormalities among the study population was 3% for FV Leiden and 1.7% for prothrombin gene 20210A. One patient was heterozygous for both FV Leiden and prothrombin gene 20210A. No patient with a congenital prothrombotic abnormality presented with a TE. Four of eight patients who were positive for APLA had TEs, indicating a trend toward an association (P = 0.29). Therefore, a trend was seen toward an association between the presence of APLA and TEs, but no trend was seen in patients with either congenital prothrombotic disorder.
|Variable||Thromboembolic event||No thromboembolic event|
|Factor V Leiden|
|Prothrombin gene 20210a|
TEs are serious complications in children with ALL and are associated with ASP therapy.2–14 However, the true prevalence and extent of TEs in this patient population are unknown. The PARKAA Study was an open, randomized, controlled, extended, Phase II study with the primary objective of determining the prevalence of TEs in children with ALL during treatment with ASP. A secondary objective of the PARKAA Study was to determine whether there was a trend toward an association between specific congenital and acquired prothrombotic risk factors and the development of TEs. The results of the PARKAA Study show that the prevalence of TEs in children with ALL receiving ASP is exceedingly high at 36.7% (95%CI, 24.4–48.8%), and that the extent of occlusion is likely clinically significant. Second, a trend was seen toward an association between APLA and TEs. However, no trend was seen toward an association between TEs and FV Leiden or prothrombin gene 20210A.
In previous reports, the prevalence of TEs among children with ALL receiving ASP ranged between 1% and 14%, with the variability reflecting study design and the time the period of the reports, with prevalences between 11% and 14% in more recent, prospective studies.4, 7–13 In all previous studies, prevalences were reported based on symptomatic TEs only. The prevalence of symptomatic TEs in the PARKAA Study was somewhat lower at only 5% (95%CI, 1–14%). However, the increased overall prevalence of TEs reported in the PARKAA Study (36.7%), compared with previous studies, reflects the systematic and comprehensive evaluation of patients with sensitive radiographic tests immediately after treatment with ASP. The rigorous evaluation of the primary outcome performed in the PARKAA Study identified asymptomatic TEs with significant occlusion.
To date, the most commonly reported location of TEs in children with ALL receiving ASP was within the CNS (> 50%).4, 6, 13, 14, 34–40 In the PARKAA Study, the percentage of TEs within the CNS was low (4.5%), whereas the percentage of TEs in the upper venous system was high (95.5%). The discrepancy can be explained by differences in study design. First, a majority of previous articles were case reports and case series which are susceptible to a reporting bias publishing on symptomatic TEs within the CNS. In the PARKAA Study, nonselected patients were followed prospectively throughout the study period. Second, in the PARKAA Study, there was a systematic assessment using accurate objective tests. Therefore, the location of TEs reported from the PARKAA Study represents the least biased data available.
The vast majority of TEs in the PARKAA Study were asymptomatic and were not associated with the classic clinical symptoms of TEs in adults: edema, pain, and skin discoloration. The development of the latter clinical symptoms of TEs likely reflects the usually lower limb proportion, extensiveness of the TEs, and acuteness of venous obstruction. In the PARKAA Study, the majority of TEs also were extensive, with 33% of TEs occluding > 75% of the greatest vessel dimension. However, the TEs were located in the upper central venous system and likely occurred gradually in response to the continuous use of ASP and the presence of a CVL. In addition, collaterals were present in 60% of patients with TEs and were classified as major in 40%. The development of significant collaterals likely minimized the acute symptoms of arm, neck, or facial swelling while the upper deep venous system was being destroyed gradually and silently. In addition, in three children, the thrombosis extended into the right atrium, placing them at risk for pulmonary embolism.41 Therefore, in children with TEs in the upper system, major destruction of the venous system and a real risk for pulmonary embolism were observed in the presence of asymptomatic TEs. However, the long-term morbidity related to these TEs needs to be determined in carefully designed, long-term follow-up studies in this study population.
Studies in adults have suggested that asymptomatic TEs may not be clinically important. However, these studies cannot be extrapolated to findings in children because of both differing etiologies and locations of the TE. In adults, TEs mostly are idiopathic and occur predominantly in the lower limb. In children, TEs occur in the presence of one or more risk factors and occur primarily in the upper venous system. The reasons why asymptomatic TEs are of clinical importance in children include the persistent presence of an ongoing stimulus for propagation of the clot until acute symptoms become evident, such as PE, occlusion or rupture of collateral varices, repeated infection, repeated loss of catheter patency requiring local thrombolytic therapy or replacement which necessitates a general anesthetic, and postthrombotic syndrome which may require several years to become apparent and may cause considerable long-term morbidity.41–43 Data from the PARKAA Study support two of these clinical observations. First, there was a trend toward more CVL-related infections in children who had TEs compared with children who did not have TEs. Second, there was a trend toward more UK use in patients who had TEs compared with patients who did not have TEs. In addition, one patient who had a TE required CVL replacement.
Mechanisms that have been postulated as responsible for TEs in patients with childhood ALL include the disease itself, treatment with ASP, and congenital or acquired prothrombotic risk factors.2, 9 The rationale for the latter mechanism is related to two publications in the literature reporting that, in this population, patients with congenital prothrombotic risk factors have an increased risk for TEs.10, 11 However, another study found no association.12 A secondary objective of the PARKAA Study was to determine whether there was a trend toward an association between FV Leiden, prothrombin gene 20210A, and/or APLA and patients at high risk for TEs. Other mechanisms have been hypothesized in the literature, including decreased protein C, protein S, and plasminogen. Due to the limited number of patients, only a selection of the proposed mechanisms could be assessed, and the evidence for the markers tested were the most convincing. Because only limited numbers of patients were assessed, the study was underpowered to detect a statistically significant association. The overall prevalence of FV Leiden and prothrombin gene 20210A in the study population was comparable to what has been reported in the literature.25, 26, 28–30, 44, 45 Neither of these two prothrombotic disorders showed the ability to be predictive of children with TEs. However, a trend was seen toward an association between APLA and TEs. These findings support the hypothesis that the risk for TEs is related overwhelmingly to acquired disorders, and/or iatrogenic factors, and/or the disease process. Larger cohort studies need to be performed to test this hypothesis.
The following potential selection biases in the study patient population should be noted. The overall degree of the effect of these biases cannot be stated with certainty, although the direction of the effect can be predicted with reasonable confidence. First, a bias may have been introduced because the study was designed for the use of a clinical product. For example, one exclusion criterion was any medical condition that may have interfered with the assessment of study medication. Patients who were excluded because they fulfilled such a criterion were likely to be more ill compared with patients who did not fulfill the criterion. Second, another possible selection bias was related to the fact that patients may have refused to enter the study because of possible treatment with a plasma-derived product and multiple, invasive radiographic tests, such as bilateral venograms. Again, it is reasonable to hypothesize that patients who refused to participate because of blood product usage or radiographic tests may have been more ill and, thus, less likely to agree to participate in a clinical study. Therefore, both of these biases lean toward the inclusion of a less complicated patient population. Patients with more serious complications are likely at an increased risk for TEs. Therefore, the potential effect of these biases is that the reported prevalence may be an underestimate.
In summary, The PARKAA Study provided cogent evidence showing that the prevalence of TEs in children with ALL who are receiving ASP is high and that the extent of TEs is likely clinically significant. Secondly, it was found that the presence of congenital prothrombotic markers did not show a trend to predicting patients who were at increased risk. However, a trend was seen toward an association between the presence of APLA and TEs. These results provide a rationale for future studies assessing primary prophylaxis in children who are receiving ASP.
The authors acknowledge and thank the study coordinators: Patsy Vegh, Monica Adams, Karen Bilynsky, Alex Blay, Tracy Corr, Elaine Dollard, Christine McDonald, Julie Nichols, Judy Powers, Chris Tremblay, and Hanna Zaire.
- 2Thrombosis in children with acute lymphoblastic leukemia with special regard to asparaginase treatment. Hemostaseologie (Germany). 1992; 12: 35–43., .