Real-world analysis of strategies to prevent thrombosis and bleeding in adults with ALL treated with asparaginase Open Access

The prioritization of pediatric-inspired regimens for the treatment of adolescents and young adults (AYAs) with acute lymphoblastic leukemia (ALL) has resulted in improved outcomes.^1,2 Higher and more frequent dosing of asparaginase (Asp) is a key component of these pediatric-inspired regimens.³ Despite recent advances in the treatment of adult ALL using targeted therapies, Asp remains an integral component of frontline regimens for both the AYA and older adult populations.^4-7 Approval of the PEGylated form of Asp, PEGasparaginase (PEG-Asp), for first-line treatment of newly diagnosed pediatric ALL, occurred in 2006, allowing for dosing once every 14 days vs 3-times-weekly dosing with native Escherichia coli Asp.⁸ This significantly improved the ease of administering Asp therapy, and PEG-Asp has become the preferred form of Asp therapy in adult patients with ALL.

Asp is associated with significant and unique toxicities, including hypersensitivity reactions, pancreatitis, thrombosis and bleeding, elevated liver enzymes, hyperglycemia, and hypertriglyceridemia.^9,10 Asp-associated venous thromboembolism (VTE) results from the depletion of antithrombotic factors, especially antithrombin (AT), and can be particularly problematic in the adult ALL population for whom this risk increases with advanced patient age.^11-13 Although the reported rate of Asp-associated VTE among adults varies significantly from study to study, it has been reported to be as high as 34% and increasing to 42% among adults aged ≥30 years.¹⁴ A recent meta-analysis of adult Asp-associated VTE reported an overall event rate of 15.8% and higher rates of VTE among patients receiving PEG-Asp than native E coli Asp.¹⁵ When Asp-associated VTE occurs, treatment is with low-molecular-weight heparin (LMWH), and continuation of Asp therapy after thrombotic events is likely safe without significant risk for recurrent VTE.^9-11,14,16 

Strategies for preventing Asp-associated VTE include AT repletion and chemoprophylaxis with LMWH; however, data supporting their efficacy in adults are limited. A Cochrane systemic review from 2020 evaluating prophylaxis of Asp-associated VTE in adults was unable to provide recommendations regarding the use of thromboprophylaxis due to a lack of high-quality evidence.¹⁷ The randomized, controlled PAARKA trial in the pediatric ALL population showed a trend toward a reduction in Asp-associated VTE for children receiving AT repletion vs no repletion, but there have been no randomized studies of AT prophylaxis for Asp-associated VTE in adults with ALL.¹⁸ The nonrandomized, retrospective studies that have been performed in adults show conflicting results, with most of these studies having small sample sizes.^19-26 As with the data, society and expert panel guidelines contradict each other on whether to recommend prophylactic AT replacement when AT levels drop below a prespecified threshold.^9-11 Likewise, evidence for Asp-associated VTE prophylaxis with LMWH in adults with ALL is also limited. Prospective studies in children have shown a benefit of chemoprophylaxis with LMWH during ALL treatment.^27,28 In adult ALL, small single-center evaluations of prophylactic LMWH have shown conflicting results.^29-35 However, none of these studies reported a significant increase in the rate of major bleeding with the addition of LMWH. Acknowledging these limited data, expert recommendations generally favor LMWH prophylaxis during Asp-containing ALL therapy in the absence of severe thrombocytopenia (typically platelet count of <30 × 10⁹/L) but they fail to provide a recommendation on dosing.^9-11 

Further complicating the prevention and management of Asp-associated VTE is the increased risk of bleeding during ALL treatment, which can be attributed to thrombocytopenia and hypofibrinogenemia. In addition to depleting AT levels, Asp depletes fibrinogen with rates of severe hypofibrinogenemia (fibrinogen of <100 mg/dL) reported in the 40% to 50% range in studies using PEG-Asp.^4,16 Patients have been given cryoprecipitate or fibrinogen concentrate for severe hypofibrinogenemia during ALL treatments, but several studies observed an increased rate of VTE among patients receiving these thrombogenic products.^16,20 Therefore, expert guidance is to administer cryoprecipitate or fibrinogen concentrate only in patients with active bleeding and to use a conservative repletion threshold.^9,10 

Given the lack of high-quality evidence evaluating prophylactic strategies for Asp-associated VTE in adults, we performed a multicenter retrospective analysis of Asp-associated VTE rates and prophylactic strategies among adults treated for ALL using Asp-containing regimens.

This multicenter retrospective study analyzed the medical records of adults with ALL who received an Asp-containing induction regimen at 6 academic medical centers in the United States between February 2010 and October 2021. Patients with Philadelphia chromosome–positive and –negative B-cell ALL, T-cell ALL, natural killer cell leukemia, and mixed-phenotype acute leukemia were permitted. Institutional review board approval was obtained for participation in this study from each study site. Data, including demographic information, VTE and bleeding risk factors, ALL disease characteristics, chemotherapy received, and baseline laboratory data, were retrospectively collected.

Two of the 6 study sites had protocols for AT monitoring and repletion during the study period. AT levels at the 2 sites were evaluated at baseline and 2 to 3 times weekly after PEG-Asp administration while the patients remained hospitalized. The threshold for prophylactic AT repletion at 1 of the sites was <60%, with a repletion goal of 100%. The other site used an AT repletion trigger of <50% and repleted to a goal of 80%. AT monitoring was discontinued 4 weeks after Asp administration or until normalization to 2 levels above the repletion threshold.

Of 6 study sites, 2 had protocols for administration of prophylactic anticoagulation among patients receiving Asp for ALL induction. One site administered 40 mg of enoxaparin subcutaneous once per day, and the other site used 40 mg of enoxaparin daily or 60 mg if the patient’s weight was ≥80 kg. The first site administered prophylactic LMWH from admission to discharge but held the medication if the patient’s platelet count decreased to <50 × 10⁹/L. Unfractionated heparin was substituted for LMWH (5000 units every 8 hours administered subcutaneously) if the patient’s creatinine clearance was <30 mL/min. The second site administered LMWH for 4 weeks after PEG-Asp if the patient’s creatinine clearance was ≥30 mL/min and there was no significant bleeding. LMWH was held if the patient’s platelet count dropped to <30 × 10⁹/L. A further 2 sites had no specific protocol for prophylactic anticoagulation and deferred this decision to provider preferences. The remaining 2 study sites did not administer prophylactic anticoagulation to any patients receiving Asp.

The primary objective of this study was to evaluate whether VTE prophylaxis with AT supplementation or anticoagulation led to a difference in the rates of symptomatic thrombosis or bleeding during the 28-day ALL induction course compared with no prophylaxis. Only index venous thrombotic events were included in our analysis and described by regional location and whether they were catheter associated.³⁶ Only bleeding events that qualified as major or clinically relevant nonmajor bleeding (CRNMB) as defined by the International Society on Thrombosis and Haemostasis criteria were included in our analysis.³⁷ Secondary outcomes evaluated in our study included the impact of patient- and treatment-specific factors on VTE and bleeding rates among the study participants.

Standard descriptive statistics were used to provide a summary of the patient characteristics. Logistic regression was used to model binary outcomes, including induction of thrombosis and bleeding. The Firth penalized likelihood approach was used to stabilize the model estimates because of sparse data.³⁸ For the logistic regression analyses, model covariates included patient characteristics, type of steroid in induction, VTE and bleeding risk factors, type of ALL, type of central venous catheter (CVC), prophylaxis strategy, and laboratory counts. Univariate models were first evaluated to build multivariate models. Only covariates with a P value of <.10 were considered for the final binary outcome models. Once combined in a multivariate model, covariates were sequentially removed by the highest P value if they were no longer significant in the presence of other covariates at α = 0.05. An analysis was performed to evaluate the relationship between thrombosis and the cumulative dose of Asp, which allowed for the assessment of a possible cutoff point across the range of this variable. Five patients were missing data for thrombosis and bleeding events and were removed from the logistic regression models. All analyses were performed using SAS (version 9.4; SAS Institute, Cary, NC).

A total of 233 patients received an Asp-containing induction regimen for ALL at the participating institutions. The median age of our study population was 33 years, with ages ranging from 18 to 78 years. Ninety-four percent of patients had newly diagnosed ALL, with B-cell and T-cell ALL representing 67.4% and 30.9% of the cohort, respectively. Additional baseline demographic information and disease characteristics are presented in Table 1.

PEG-Asp was the most common Asp formulation administered to 98.3% of patients in our study. Eighty-eight percent of the patients received Asp as an IV infusion and 10.7% received an intramuscular injection. Ninety-eight percent of the patients received a single dose of Asp, and 6.4% received 2 doses. Most Asp doses were 2000 or 2500 IU/m² (78.1%); however, 17.6% of the patients received a lower dose of 1000 IU/m². The dose of Asp was capped for 9.4% of patients in our cohort (most commonly at 3750 IU for 18/22 patients). Asp was not capped for 74.2% of the cohort and data on dose capping was missing for 16.3%. The induction chemotherapy regimen, steroid formulation, and type of CVC used during the induction course in our study population are described in Table 2. Fifty-eight patients (24.9%) in our study population received at least 1 dose of AT, 41 patients (17.6%) received subcutaneous LMWH or unfractionated heparin chemoprophylaxis, and 21 patients (9.0%) received both. The median duration of anticoagulation during induction was 10 days (interquartile range, 6.5-18.5). Cryoprecipitate transfusions were administered to 145 patients (62.2%), with a fibrinogen trigger for transfusion of 100 to 150 mg/dL in 10.7%, 50 to 99 mg/dL in 80.0%, and <50 mg/dL in 2.1% of patients who received transfusions. Fresh frozen plasma (FFP) transfusion was administered to 22 patients (9.4%).

Thirty-two patients (13.7%) experienced symptomatic venous thrombotic events. Four patients had multiple sites of thrombosis at index event and 17 VTE events were line-associated (52.7%). The most common sites of thrombosis were upper extremity deep vein thromboses (61.1%), followed by pulmonary emboli (16.7%), lower extremity deep vein thromboses (11.1%), cerebral venous sinus thrombosis (CVST; 5.6%), and jugular vein thrombosis (5.6%; Table 3). Twelve patients (5.2%) experienced CRNMB (4 patients) or major bleeding (8 patients) during induction. Table 3 lists the bleeding location and severity. Thrombotic events occurred at a median of 14 days (range, 3-26) and bleeding events occurred at a median of 13.5 days (range, 0-24) after the first dose of Asp. There was no statistically significant association between the cumulative Asp dose received and thrombosis.

The logistic regression modeling results for induction thrombosis and bleeding are shown in Table 4 and Figure 1. The odds of thrombosis were not different between patients who did or did not receive prophylaxis with AT repletion or anticoagulation (odds ratio [OR], 0.734; 95% confidence interval [CI], 0.304-1.773; P = .49 for AT repletion; and OR, 1.275; 95% CI, 0.490-3.316; P = .62 for prophylactic anticoagulation). Similarly, AT monitoring did not affect the odds of developing thrombosis (OR, 1.076; 95% CI, 0.509-2.274; P = .85). The odds of thrombosis were not increased based on patient demographics, medical comorbidities, B-cell vs T-cell ALL, type of steroid administered during induction, menstrual suppression with leuprolide, transfusion of cryoprecipitate, or FFP. However, we observed a statistically significant increase in the odds of thrombosis among patients with peripherally inserted central catheters (PICC) compared with among patients with other types of CVCs (OR, 4.112; 95% CI, 1.622-10.427; P = .01). When excluding line-associated thromboses from our analysis, this increased odds of thrombosis among patients with PICCs was no longer seen (supplemental Table 1).

The odds of bleeding during induction did not differ between patients who did or did not receive prophylaxis with AT repletion or anticoagulation (OR, 0.338; 95% CI, 0.059-1.934; P = .22 for AT repletion; and OR, 0.165; 95% CI, 0.009-2.947; P = .22 for prophylactic anticoagulation). Similarly, the odds of bleeding did not differ according to patient demographics, medical comorbidities, administration of therapeutic anticoagulation during induction, platelet count or fibrinogen nadir, or transfusion of cryoprecipitate or FFP. There was a statistically significant increase in the odds of bleeding in patients who received menstrual suppression during induction (OR, 5.084; 95% CI, 1.034-24.997; P = .05) or who had active infection during induction (OR, 5.084; 95% CI, 1.034-24.997; P = .05).

Supplemental Tables 1 and 2 show the disease status at the end of induction, postinduction treatment courses received, and patient status at the last follow-up in the patient cohort.

Asp-associated thrombosis remains a significant complication of current treatment approaches for adult ALL, and the optimal strategy for preventing this complication has not been determined. In our multicenter, retrospective study evaluating the experience of 233 adult patients treated with conventional Asp-containing protocols, we identified a thrombotic event rate of 13.7%, which is consistent with recent studies evaluating thrombosis among patients receiving PEG-Asp and the formulation of Asp received by 98.3% of the patients in our study.^15,16 In multivariate analysis, the odds of developing a thrombotic complication did not differ regardless of whether the patients received AT monitoring, prophylactic AT repletion, or prophylactic anticoagulation. Similarly, the cumulative dose of Asp received during induction did not have a statistically significant impact on the rate of thrombosis in our study.

Approximately half of the venous thrombotic events in our study (53.1%) were associated with central lines. CVCs have been identified as a risk factor for VTE among patients with ALL receiving Asp.¹³ Indeed, the type of CVC was the only variable with a statistically significant difference in the odds of induction thrombosis among our patient cohort, with the odds of thrombosis being 4.1 times higher for patients with PICC than for those with CVC, and, when removing the line-associated thromboses from our analysis, the type of venous access was no longer associated with increased odds of VTE. An increased risk of VTE with PICCs compared with CVCs has been reported in patients with various types of solid organ and hematologic malignancies, and our data indicate that this holds true among the adult ALL population.^39,40 CVST is a rare but serious complication described in patients receiving Asp-based regimens.⁴¹ CVST occurred in 2 patients in our study (0.9% of the overall cohort), accounting for 5.6% of all VTE events identified. One of the 2 patients had associated central nervous system hemorrhage, and neither received prophylactic AT repletion or anticoagulation therapy.

Fifty-eight patients in our study received at least 1 dose of prophylactic AT, which was 24.9% of the overall study population, and 60.4% of the 96 patients treated at the 2 centers that had protocols for AT monitoring and repletion. The lack of a statistical difference in the odds of thrombosis among the patients who were monitored with AT levels or received prophylactic AT repletion observed in our study is consistent with the results of a retrospective analysis of 75 adult patients with ALL treated with PEG-Asp between 2014 and 2017.¹⁹ In this study, the incidence of VTE events did not differ between patients with AT monitoring and repletion for AT levels of <60% (n = 47) and patients without prophylactic AT repletion (n = 28): 17% in patients with AT repletion vs 11% in the control group (P = .52). Another evaluation of prophylactic AT repletion from the GRAALL-2005 study of 784 patients treated with native E coli L-Asp also failed to show a benefit for AT repletion, which was administered to 87% of the patients in the analysis (VTE incidence of 8% with AT repletion vs 14% without; P = .10).²⁰ 

Several small retrospective studies have reported the benefits of prophylactic AT repletion in reducing the rate of Asp-associated thrombosis in adults with ALL.^21-25 These studies were limited by the use of historical controls, for which the rates of VTE were compared before and after the implementation of AT monitoring and repletion protocols, and small sample size. The largest retrospective analysis showing the benefit of AT repletion is the multicenter CAPELAL study of 214 patients published in 2008, which showed a significant reduction in E coli L-Asp–associated VTE rates among patients receiving prophylactic AT repletion (4.8% vs 12.2%; P = .04).²⁶ This finding is notable given the large number of patients who received AT repletion (41% of the cohort). Although our study is also vulnerable to the confounding inherent in all retrospective studies and only 24.9% of patients received AT repletion, it has the benefit of including a larger study population than most of the recent studies evaluating the use of PEG-Asp because of its multicenter design and a contemporaneous cohort of patients who did not receive AT repletion instead of using a historical control.

Prophylactic anticoagulation therapy, predominantly LMWH, was administered to 41 patients (17.6%) in our study. Multivariate analysis showed no difference in the odds of thrombosis, regardless of whether the patients received prophylactic anticoagulation. In several retrospective studies, prophylactic anticoagulation administered during induction has not been shown to be beneficial to adults receiving Asp.^29-32 However, most of these studies included a very small number of adults and were underpowered to detect differences in outcomes. Additionally, prophylactic anticoagulation is held when platelet counts decrease below a prespecified threshold (typically <30 × 10⁹/L or 50 × 10⁹/L), resulting in a limited duration of therapy throughout induction when cytopenias are most prevalent, which may mitigate potential antithrombotic benefits. The median duration of prophylactic anticoagulation during induction in our study was also short at just 10 days. A benefit of prophylactic anticoagulation has been shown when administered with Asp-based therapy during the intensification phase of ALL therapy.^33,34 In addition, a meta-analysis incorporating many of the smaller studies, published in 2023, showed a decreased incidence of VTE among patients receiving LMWH during Asp-based therapy.¹⁵ Therefore, the lack of benefit observed in our analysis may be because our cohort was underpowered to detect a difference in thrombosis rates with and without prophylactic anticoagulation. Similarly, our study may have been underpowered to detect the differences in previously described risk factors for Asp-associated thrombosis including age, obesity, prednisone vs dexamethasone use, and cryoprecipitate or FFP transfusion.^{13,14,16,20,33} 

A recently published single-center, retrospective study reports on a novel strategy using intermediate-dosing enoxaparin (1 mg/kg per day) for AYA patients with ALL receiving PEG-Asp induction regimens.³⁵ Of 62 patients receiving this higher dosing of LMWH, only 4 experienced VTE events (6.4%), 3 of which were catheter-associated events. To maximize the duration of prophylactic anticoagulation, patients were transfused to maintain platelet counts of ≥30 × 10⁹/L. These results raise the question of whether the prophylactic dosing of anticoagulation in our and prior studies and the short duration of therapy because of thrombocytopenia may have mitigated a potential benefit to prophylactic anticoagulation following the strategy above. Prospective, randomized studies are required to confirm the reproducibility of these results.

Twelve patients (5.2%) in our study experienced CRNMB or major bleeding events during their induction course, with major bleeding occurring in 8 patients (3.4%). This is consistent with the rates of major or grade ≥3 bleeding reported in the literature among adult patients with ALL treated with Asp-based regimens, which range from 0% to 5%.^20,30-34 These studies did not show an increased rate of significant bleeding among patients receiving prophylactic anticoagulation. Similarly, in our multivariate analysis, the odds of bleeding during induction did not increase among patients who received prophylactic anticoagulation, and none of the patients with major bleeding received therapeutic or prophylactic anticoagulation during their induction course (Table 4).

Of 15 patients who received menstrual suppression with gonadotropin releasing hormone analogues in our study, 4 experienced vaginal bleeding events although none of these were classified as major bleeding. This explains the 5-times increased odds of bleeding during induction that was associated with the receipt of menstrual suppression in our multivariate analysis as a result of increased baseline risk of bleeding in premenopausal women rather than being attributed to the intervention itself (Table 4). The odds of bleeding were also increased among patients with active infection, which could be indicative of a more complicated induction course overall.

The limitations of our study include its retrospective nature and the small sample size. Although it is one of the larger retrospective studies evaluating the role of prophylactic AT repletion and anticoagulation in 233 patients receiving Asp overall, only 58 (24.9%) patients received AT repletion and only 41 (17.6%) received prophylactic anticoagulation. Additionally, all patients who underwent AT monitoring and repletion throughout their induction course were treated at 2 of 6 centers participating in the trial. Variations in institution-specific practices between study sites could introduce confounding factors and bias into our results. Twenty-one patients in our cohort received both AT repletion and prophylactic anticoagulation. Having a large portion of patients receiving both interventions adds confounding when interpreting the impact of either intervention.

In conclusion, our multicenter retrospective study of 233 adult patients with ALL receiving Asp-based induction therapy showed no difference in the odds of thrombosis with prophylactic AT repletion or prophylactic anticoagulation, with a thrombotic event rate of 13.7% overall. CRNMB and major bleeding occurred in 5.2% of patients (3.4% with major bleeding), with no difference in the odds of bleeding with prophylactic or therapeutic anticoagulation or transfusions of cryoprecipitate. Thrombosis was more likely among patients with PICC lines than among those with CVCs, and bleeding was more likely in patients receiving menstrual suppression or who had active infection during their induction course. Large prospective randomized trials are needed to determine whether prophylaxis with AT repletion and anticoagulation decreases thrombosis among adults with ALL receiving PEG-Asp without increasing bleeding with special attention to novel anticoagulation dosing strategies.

Contribution: J.F.M. and F.E.C. contributed to the study conception and design; D.B. and B.H. analyzed the data; J.F.M. wrote the first draft of the manuscript with input from F.E.C.; J.F.M. and F.E.C. accessed and verified the data; and all authors collected and interpreted the data, and edited and approved the manuscript.

Conflict-of-interest disclosure: R.M. is an employee at Vanderbilt University Medical Center and was employed by Pfizer and held stock in the company at the time of publication. B.D. has served on consulting/advisory boards for Janssen, Pluri Biotech, Boxer Capital, Ellipses Pharma, Lumanity, Autolus, Acrotech, ADC Therapeutics, and Gilead Sciences; and received institutional research funding from Janssen, Angiocrine, Pfizer, Poseida, MEI Pharma, Orca Bio, Wugen, AlloVir, National Cancer Institute, Atara, Gilead Sciences, Molecular Templates, Bristol Myers Squibb (BMS), AstraZeneca, and Adicet. H.W. works for an institution that receives a grant or contract from Schrödinger Inc. S.C.L. received research funding from AbbVie, BMS, GlaxoSmithKline (GSK), and Jazz Pharmaceuticals. R.M.S. has served in a consulting or advisory role for, and received honoraria from, BMS, Curio Science, Gilead Sciences, Kura Oncology, Rigel, and Servier; and serves on a steering committee for Servier. V.H.D. received a travel grant from DAVA Oncology. T.W.L. has received honoraria for consulting/advisory boards from AbbVie, Agilix, Agios/Servier, Apellis, Astellas, AstraZeneca, BeiGene, Blue Note, BMS/Celgene, Genentech, GSK, Lilly, Meter Health, Novartis, and Pfizer; has received speaking-related honoraria from AbbVie, Agios, Astellas, BMS/Celgene, Incyte, and Rigel; has equity interest in Dosentrx (stock options in a privately-held company); receives royalties from UpToDate; and reports research funding from the AbbVie, American Cancer Society, AstraZeneca, BMS, Deverra Therapeutics, Duke University, GSK, Jazz Pharmaceuticals, Leukemia and Lymphoma Society, National Institute of Nursing Research/National Institutes of Health, and Seattle Genetics. F.E.C. is a consultant for SPD Oncology, Amgen, CTI BioPharma, AbbVie, MorphoSys, Association of Cancer Care Centers, and PharmaEssentia; has received clinical trial grant support (principal investigator) to the University of Virginia from Amgen, BMS, Celgene, SPD Oncology, Sanofi, FibroGen, PharmaEssentia, BioSight, MEI Pharma, Novartis, and Arog Pharmaceuticals; and has received a travel grant from DAVA Oncology. The remaining authors declare no competing financial interests.

Correspondence: Firas El Chaer, Division of Hematology and Oncology, University of Virginia Comprehensive Cancer Center, West Complex Room 6248, 1300 Jefferson Park Ave, PO Box 800716, Charlottesville, VA 22908; email: fe2gh@uvahealth.org.