Thrombosis and acute leukemia

Abstract

Thrombosis is a common complication in patients with acute leukemia. While the presence of central venous lines, concomitant steroids, the use of Escherichia coli asparaginase and hereditary thrombophilic abnormalities are known risk factors for thrombosis in children, information on the pathogenesis, risk factors, and clinical outcome of thrombosis in adult patients with acute lymphoid leukemia (ALL) or acute myeloid leukemia (AML) is still scarce. Expert consensus and guidelines regarding leukemia-specific risk factors, thrombosis prevention, and treatment strategies, as well as optimal type of central venous catheter in acute leukemia patients are required. It is likely that each subtype of acute leukemia represents a different setting for the development of thrombosis and the risk of bleeding. This is perhaps due to a combination of different disease-specific pathogenic mechanisms of thrombosis, including the type of chemotherapy protocol chosen, the underlying patients health, associated risk factors, as well as the biology of the disease itself. The risk of thrombosis may also vary according to ethnicity and prevalence of hereditary risk factors for thrombosis; thus, it is advisable for Latin American, Asian, and African countries to report on their specific patient population.

Keywords:

• Acute leukemia
• Thrombosis
• Risk factors

Thrombosis and Cancer

Thrombosis is a common complication in cancer patients and the association between these two entities has been known for centuries. Deep venous thrombosis and pulmonary embolism (PE) are the most common cancer-associated thrombotic events. Cancer patients account for 20% of all patients with venous thromboembolism (VTE), and in recent years, the reported incidence has been as high as 28% of hospitalized cancer patients.1

The mortality rate in patients with cancer and thrombosis is higher (16·3%) than in those without thrombosis (6·3%), particularly in patients who develop PE (24·8%).2 Some tumor types are more likely to be associated with the development of thromboses: pancreas, brain, ovary, and lung. About 10% of patients with idiopathic VTE will be diagnosed with cancer within the following 10 years (more than 75% will be diagnosed within the first year).3

A validated scoring system predicting the development of VTE in cancer patients has been proposed, and is appropriate for solid tumors;4 yet, its parameters such as high leucocyte or platelet counts, are not useful for patients with acute leukemia.

Prophylaxis with low-molecular-weight heparin (LMWH) or unfractionated heparin is indicated in patients with cancer undergoing major surgery. This same prophylactic measures may benefit medical oncology patients admitted to the hospital with an acute illness.3 However, prophylaxis may represent a risk per se in patients with acute leukemia with very low platelet counts or with coagulation disorders.

Initial treatment of VTE consists of anticoagulant therapy; LMWH or unfractionated heparin are the preferred frontline drugs. Secondary prophylaxis with vitamin K antagonists can also be initiated on the same day as heparin therapy. Treatment must be adjusted to maintain an international normalized ratio within 2·0–3·0. The annual incidence of recurrent VTE in patients without cancer is 8%, but in patients with cancer, this rate increases two- to threefold. This could be a consequence of the difficulty in maintaining an optimal international normalized ratio in these patients, due to drug interactions, anorexia, malnutrition, liver dysfunction, and poor gastrointestinal tolerance,3 as well as the presence of tumor-related prothrombotic factors such as mechanical vascular compression or an increase in molecular activators of coagulation due to the overexpression of tissue factor by malignant cells.

The Italian Society for Haemostasis and Thrombosis recommends LMWH for the first 6 months in patients with VTE and hematological malignancies. In patients with severe and prolonged thrombocytopenia, the use of LMWH is preferable to oral anticoagulant therapy.5 Recommendations for treatment of thrombosis in patients with cancer are mostly based on studies of patients with solid tumors.3,5,6 Guidelines for prophylaxis and management of VTE based on prospective studies conducted in acute leukemia patients are lacking.

Thrombosis and Acute Leukemia

Patients with acute leukemia present both an increased risk of hemorrhage, as well as of thrombosis. The incidence of thrombosis varies between 2 and 36%.7^–14

Relevant aspects regarding patients with acute leukemia and thrombosis are as follows:

1. association between central catheters and thrombosis and its clinical variability on presentation (symptomatic or asymptomatic);

2. association between L-asparaginase during induction chemotherapy and thrombosis in patients with acute lymphoid leukemia (ALL);

3. recognition of thrombosis as a frequent and perhaps underestimated complication in patients with acute promyelocytic leukemia (APL);

4. inconsistencies among reports in the literature regarding risk factors for thrombosis (gender, age, leukemia subtype, catheter characteristics, comorbidity, and genetic mutations) and their impact on overall survival.

ALL

As with any other disease, patients with ALL develop thromboses as a result of an interaction of factors. According to current evidence, the main contributors include the disease itself, the type of administered chemotherapy, central venous catheters (CVC), genetic abnormalities, as well as an acquired predisposition. Most published studies are in pediatric patients. As reported by Athale and Chan,15 the average incidence of thrombosis in children with ALL is 3%. According to a prospective study in adults with ALL, the incidence of thrombosis increases to 9·6%.10

Variation in the incidence of thrombosis depends on several factors, such as the study design, prospective versus retrospective,16,17 whereby the former tend to report higher incidences of thrombotic phenomena. Furthermore, studies designed to detect asymptomatic thrombosis have also reported increased thrombosis incidence rates.15 The incidence of thrombosis is also related to the treatment regimen.

The German group has reported different incidence rates of thrombosis when comparing patients receiving the Berlin–Frankfurt–Münster (BFM) or the Cooperative Study Group for Childhood Acute Lymphoblastic Leukemia (COALL) chemotherapy protocols. Patients receiving the COALL protocol had a 0·8% incidence of thrombosis during induction chemotherapy, whereas patients receiving the BFM protocol had an incidence of 10%. Several risk factors were evaluated in the multivariate analysis. Only the concomitant administration of Escherichia coli asparaginase/prednisone to leukemic children with a prothrombotic risk factor was found to increase the risk of thrombosis (odds ratio: 34·5; 95% confidence interval: 4·39–271·42; P = 0·0008). Interestingly, the differences in thrombosis rates were observed only during induction chemotherapy.18

The incidence of thrombosis also varies according to the study period (before or after 1990). The overall incidence of VTE in studies conducted or reported before or after 1990 is 1·8 and 4·7%, respectively. This may be possibly explained by improved diagnostic methods, different chemotherapy regimens or an increase in suspicion and confirmation of the diagnosis of thrombosis.15 Regarding the site of thrombosis, a reported 52% of patients with central nervous system thromboembolism have venous sinus involvement. Thrombosis may affect overall survival, quality of life, and cognitive function. There are no data regarding the recurrence of thrombosis in this population. Patients are reported to have residual neurological deficits or seizure disorders. Associated morbidity is approximately 15–20% in cases of central nervous system (CNS) thrombosis.15

L-asparaginase decreases plasminogen, fibrinogen, and antithrombin, resulting in impaired thrombin inhibition which may contribute to the asparaginase-related dose-limiting toxicity.19-21 Changes in antithrombin (AT) and fibrinogen during induction chemotherapy with L-asparaginase were reported in a retrospective study of 214 adult patients with ALL.22 The median AT levels decreased from 120 to 59% after the fourth L-asparaginase infusion. AT levels below 60% were found in 50% of cases. Fibrinogen levels decreased from 2·9 g/l at diagnosis, to 1·9 g/l before the first infusion, they continued to decrease after the first 10 days of induction therapy and reached a median value of 1·1 g/l at the time of the fourth L-asparaginase infusion. Infusion of AT concentrate was followed by a significant increase in AT levels, from 61 to 88%. Fibrinogen levels significantly increased after fibrinogen concentrate administration, from 1 to 1·4 g/l. Fresh frozen plasma was not effective in significantly ameliorating AT or fibrinogen levels. The incidence of thrombosis was 9·8% during induction therapy and all thrombotic events occurred within 2–35 days after the first injection of L-asparaginase during induction chemotherapy. No mention of catheter-related thromboses was made, but 25% of thrombotic events developed in the upper limbs. The use of oral contraceptives was more frequent in women with thrombosis than in those without. Other factors such as familial thrombophilia, previous thrombosis, age, AT levels <60%, fibrinogen levels <0·5 g/l and low doses of heparin were similar between groups of patients with and without thrombosis. The complete remission rate was similar in patients with and without thrombosis, but thrombus development was associated with a decreased median overall survival (19 months versus 53 months), as well as a decreased disease-free survival (14 months versus 58 months).

Some studies have reported a genetic prothrombotic predisposition as an important host factor in the development of VTE in children with ALL,23 whereas others have failed to demonstrate such an association.24 Whether to administer primary anticoagulant prophylaxis with LMWH to children with ALL during induction chemotherapy, remains controversial. A recent retrospective study of 80 children with ALL,25 demonstrated an incidence of genetic thrombophilia of 22·5% (factor II G20210A and factor V Leiden). These patients received prophylactic enoxaparin, but 7·5% developed thromboembolic events. VTE was not detected in patients with factor V Leiden, suggesting that ALL patients with the PT gene mutation are at increased risk of developing clotting complications in comparison with those harboring the factor V Leiden mutation.

All patients with acute leukemia require placement of a CVC for cytotoxic chemotherapy administration. All types of CVCs are associated with infection and thrombosis. Actually, thrombosis is a risk factor for developing infection of the CVC and this association has been documented for over 25 years.26,27 Autopsy and venographic studies have demonstrated that soon after the insertion of a catheter, practically all will develop a fibrin sheath.28,29

Different diagnostic methods have been used to demonstrate that the aforementioned fibrin sheaths are always colonized by cocci.30^–32 The type of catheter-related thrombosis may be due to either clotting of the lumen (13–93%) or the development of mural thrombi in the vessel where the catheter is placed (12–74%).33^–35

Approximately, a third of CVC-related thromboses are symptomatic. Factors favoring the development of CVC-related thromboses include malignancy, thrombophilia, endothelial cell injury due to the CVC itself or to chemotherapy, the position of the catheter in the vascular system and the number of catheter lumens.34 The most important sequelae are pulmonary emboli, post-phlebitic syndrome, and infection, with an estimated frequency of 6, 15–35, and 18%, respectively.33

Acute Myeloid Leukemia (AML)

Additional risk factors for VTE in AML patients include the increased expression of tissue factor in leukemic cells, its activation on cellular surfaces, and hyperleukocytosis.36 Treatment of VTE in AML patients is challenging, because of the considerable risk of hemorrhage due to severe thrombocytopenia and coagulation and/or fibrinolysis abnormalities. A patient with bleeding and PE is not an uncommon picture for clinicians treating AML, especially at diagnosis and during the induction phase.

Thromboses are treated with LMWH. Monitoring of anti-Xa and maintenance of peak levels between 0·5 and 1 IU/ml in patients with renal failure, obesity, pregnancy, and children, as well as close observation of the platelet count are mandatory. LMWH should be decreased by 50%, if the platelet count drops to 50×10(9)/l or below, or temporarily discontinued if <20×10(9)/l.37

Important data obtained by Ku et al.11 in a population-based cohort, was used to determine the incidence of thrombosis in an American population (California) of 5394 patients with AML or ALL. The 2-year cumulative incidence of VTE was 5·2%, comparable to that in patients with solid tumors. Interestingly, 64% of thrombotic events developed within the first 3 months of the leukemia diagnosis. Risk factors for VTE in AML included female gender, older age, the number of chronic comorbidities, and the presence of an indwelling catheter. In ALL patients, the 2-year cumulative incidence of VTE was 4·5%. Risk factors for VTE included the presence of a central venous catheter, older age, and the number of chronic comorbidities. As opposed to AML patients in whom survival was not affected by VTE, ALL individuals had a 40% increase in the risk of death during the first year after diagnosis.

APL

APL patients present a particular scenario, whereby fatal hemorrhages due to disseminated intravascular coagulation were the major cause of early death before the use of all trans-retinoic acid as part of the induction treatment. Currently, the rate of fatal hemorrhages ranges between 2·4 and 6·5%.38 Increased fibrinolysis has been documented and overexpression of annexin A2 may be one of the underlying mechanisms, conditioning the hemorrhagic complications in these patients. Abnormally high levels of annexin A2 on APL cells increase plasmin generation that in turn, activates fibrinolysis. Furthermore, annexin II mRNA levels are reduced after treatment with all trans-retinoic acid39 and arsenic trioxide.40 Thrombosis is probably an underestimated complication in APL patients.41 Concomitant hemorrhage and thrombosis have also been reported during induction chemotherapy. A retrospective study of 34 consecutive APL patients in a single referral center in Israel, reported an incidence of severe thrombosis of 12%. Life-threatening bleeding occurred in 29% of patients. The most consistent hemostatic abnormality was decreased fibrinogen (<150 mg/dl) in 61% of cases. On multivariate analysis, only the leukocyte count (<30×10(9)/μl) reached statistical significance as a predictor of bleeding. Interestingly, none of the hemostatic parameters (platelet, fibrinogen, PT, and APTT) were identified as risk factors for bleeding or thrombosis. Of note, three out six patients with thrombotic events were found to be thrombophilic.42 There is a case report of a patient with APL who developed splenic, renal, and intestinal infarction due to severe acquired protein C deficiency; no evidence of disseminated intravascular coagulation was documented.43 In an Italian study,44 patients with thrombosis had a higher white blood cell count and an increased prevalence of the short PML/RARA isoform (bcr3), FLT3/ITD, CD2, and CD5 expression.

The PETHEMA study45 did not confirm these thrombosis risk factors. Instead, the identified risk factors included: fibrinogen <170 mg/dl, an M3 variant at diagnosis and tranexamic acid prophylaxis during induction chemotherapy.

Conclusions

Thrombosis is a common complication in patients with acute leukemia. Most of the information on ALL patients has been obtained from studies in children and the information in adult patients remains scant. Expert consensus and guidelines on risk factors, optimum catheter care, use and management, prevention and treatment of thrombosis are required. The role of newer anticoagulants needs to be explored in multicenter clinical trials, and risk factors should be standardized, in order to understand the disease’s pathogenesis and provide adequate treatment to these patients within a complicated setting of thrombosis and hemorrhage.

Ideally, guidelines should be based on prospectively designed studies conducted only in patients with acute leukemia. Furthermore, they should be designed separately, according to leukemia subtype (ALL, AML, and APL). It is very probable that each acute leukemia subtype represents a different physiopathogenic setting favoring the development of thrombosis.