Anticoagulant therapy in patients with cancer and thrombocytopenia

Abstract

Cancer and its treatment predispose to thromboembolic complications and thrombocytopenia. Thrombocytopenia does not protect against thromboembolism, but increases the risk of bleeding. The prognosis in venous thromboembolism (VTE) strictly depends on the efficacy of anticoagulation. The choice of anticoagulation strategy will depend on the severity of thrombocytopenia and its expected duration. Thrombotic risk is also important, determining the risk of death related to pulmonary embolism and the risk of recurrence/progression of VTE. Possible strategies include full anticoagulation and possible platelet transfusions, modification of the anticoagulation dose or interruption of anticoagulation. The review focuses on the possibilities of VTE treatment in the aspect of clinically significant thrombocytopenia with a platelet count below 50 × 10⁹/L (50 000/µl).

Graphical abstract

Defining the problem of thrombocytopenia in cancer diseases

Thrombocytopenia is defined as clinical condition with decreased blood platelet count. The degrees of thrombocytopenia according to the National Cancer Institute Common Terminology Criteria for Adverse Events (NCI CTCAE) are considered important in oncology [1] (Table 1).

Table 1.

Clinically important grades of thrombocytopenia in oncology

Platelet range (x 109/L	Grade 1	Grade 2	Grade 3	Grade 4
The upper limit	< 100	< 75	< 50	< 25
The lower limit	75	50	25

Thrombocytopenia in oncology is mainly related to either the pathogenesis of neoplastic disease or the iatrogenic effect of anticancer treatment [2] (Table 2).

Table 2.

The potential mechanisms of thrombocytopenia induced by both cancer therapies and cancer diseases

Thrombocytopenia induced by
Cancer therapies	Cancer diseases
Bone marrow suppression and destruction - Pluripotent stem cells, - Megakaryocyte progenitors, - Megakaryocytes	Bone marrow infiltration by - Cancer cells - Leukemia cells
Direct effect on platelets - Increased apoptosis - Shortened activity - Immune thrombocytopenia - Thrombotic microangiopathy	Bone marrow suppression or platelets destruction by - Cytokine-mediated mechanism - Other immune-related mechanism

Cancer itself can lead to thrombocytopenia through various specific mechanisms, such as bone marrow infiltration by cancer cells or leukemia cells. Other important mechanism may be cytokine-mediated bone marrow suppression [3].

Anticancer drugs can cause thrombocytopenia through various mechanisms, although two main pathways can be distinguished [4]. The first is bone marrow suppression, including an adverse effect on pluripotent stem cells, megakaryocyte progenitors, and megakaryocytes themselves (examples include busulfan, cisplatin, cyclophosphamide, proteasome inhibitors). The second main mechanism is related to a direct effect on platelets through a decrease in platelet kinase activity (e.g. sunitinib, dasatinib, etc.), increased platelet apoptosis (e.g. venetoclax, cisplatin), or platelet destruction (e.g. immune checkpoint inhibitors, oxaliplatin, fludarabine, gemcitabine, mitomycin C). Some oncological drugs cause pancytopenia, i.e. a decrease in the number of white blood cells, red blood cells, and platelets. However, there are oncological drugs that can selectively induce only thrombocytopenia, e.g. proteasome inhibitors [5].

Radiotherapy may induce myelosuppression [6, 7]. A mechanism is not well defined [8]. Thrombocytopenia may be an important hematologic toxicity related to radiotherapy [9].

Cancer disease and cancer therapy may determine the frequency of thrombocytopenia, the moment of its occurrence, its severity and duration. Thrombocytopenia can be a short-term and transient complication, but sometimes it can persist for quite a long time and be an irreversible complication with no chance of normalizing the platelet count.

The scale of the problem

The available data show that grade 3 or 4 thrombocytopenia (defined as platelet count < 50 000/µl) may affect 28% of patients with hematological malignancies and 6% of patients with solid tumors [10]. Even in solid tumors there are chemotherapy regimens that cause thrombocytopenia in every third patient [11].

The available data from the Beth Israel Deaconess Medical Center revealed that among 3635 cancer patients with acute thrombosis in the period between 2010 and 2021, the incidence of grade 3 or 4 thrombocytopenia (platelet count < 50 000/µl) was 30% (95% CI 27%−34%) in hematological malignancies and 7% (95% CI 6%−8%) in solid tumors [12]. Interestingly, in a group of 1514 patients after hematopoietic stem cell transplantation, the incidence of VTE was 4.6% (95% CI 3.6%−5.8%), but was significantly higher in patients experiencing thrombocytopenia: 60% of VTE occurred at a platelet count < 100 000/µl, 34% at a platelet count < 50 000/µl and 13% at a platelet count < 20 000/µl [13].The most important limitation of the study was fact that VTE prophylaxis in the form of encouraged ambulation was not standardized and clearly documented, pharmacological thromboprophylaxis remain unknown but rather it was avoided.

How to predict bleeding risk associated with thrombocytopenia?

In the COMMAND VTE registry, the incidence of major bleeding was 2.7% over 3 months among patients with cancer-related VTE, multivariate analysis showed that independent risk factors for bleeding included active cancer, age ≥ 75 years, previous major bleeding, anemia, and thrombocytopenia [14].

A platelet count below 10 000/µl is considered to be associated with the occurrence of bleeding. This platelet count is used as a criterion for prophylactic platelet transfusions in hospitalized patients undergoing hematologic stem cell transplantation or chemotherapy for hematologic malignancies or solid tumors [15].

The number of days with platelet count < 10 000/µl was significantly associated with grade 2–4 bleeding (p < 0.0001) [16]. Chemotherapy, allogeneic hematopoietic stem cell transplantation, female gender, and no prophylactic platelet transfusion were significantly associated with increased number of days with bleeding. Patients with a body temperature of 38 °C or greater had the highest risk of grade 2–4 bleeding. In this study, the World Health Organization (WHO) definition of bleeding was adopted, with grade 2–4 indicating clinically significant bleeding and grade 3–4 severe/life-threatening hemorrhage.

As described above, a platelet count < 50 000/µl is considered a risk factor for bleeding, but finally the occurrence of bleeding is more likely to be determined by an additional co-occurring factor. These may include renal dysfunction, liver disorder, hematocrit < 25%, etc. Based on the co-presence of risk factors, attempts are made to predict bleeding [17, 18]. Additional importance is played by factors directly related to the function of platelets, vascular endothelium and possible coagulation disorders [19].

In patients with a platelet count of 10–50 000/µl, the decision to transfuse platelets is based on the occurrence of bleeding [20]. ASCO guidelines recommend prophylactic platelet transfusion if platelet count < 10 000/µl and therapeutic platelet transfusion if platelet count is in the range 10–50 000/µl [21]. A major concern is the occurrence of bleeding despite platelet transfusions, which raises questions about the efficacy of this intervention [22]. Another problem is dose reduction or delay of anticancer treatment. Unfortunately, such a modification of cancer therapy means a lower relative dose intensity and may translate into a worse prognosis for survival, as observed in various cancers [23].

Is supportive treatment possible for thrombocytopenia?

Some clinicians consider the use of drugs such as tranexamic acid (TXA) and aminocaproic acid (EACA) - synthetic antifibrinolytic drugs.

EACA was rather considered for the treatment of bleeding in patients with various hematological malignancies and thrombocytopenia defined as platelet count < 20 000/µl [24–26]. One study showed the efficacy of EACA in patients with hemorrhage and thrombocytopenia in hematological malignancies (acute leukemia, non-Hodgkin’s lymphoma) and showed complete or partial cessation of bleeding in 66% and 17% of patients, which resulted in a reduction of the number of platelet and red blood cell transfusions [27].

TXA has been tried to be used to prevent bleeding and thus dependence on platelet transfusions. In two studies the efficacy of TXA included fewer bleeding episodes and fewer transfusion requirements [28, 29]. However there was a study completed without benefit [30].

Based on the available data of 11 studies evaluated 1177 patients receiving tranexamic acid or aminocaproic acid, there was no increased risk of VTE, the efficacy was expressed by a reduced risk of blood loss and the need for transfusion (RR = 0.52, 95% CI 0.34–0.80) [31].

Some randomized trials were planned among patients with hematologic malignancy undergoing chemotherapy or hematopoietic stem cell transplantation. The TREATT, an international, double-blind, phase 3 trial, revealed no significant difference in primary endpoint defined as the proportion of patients who died or had WHO grade 2 or higher bleeding: 31.7% in the tranexamic acid group vs. 34.2% in the placebo group (HR = 0.92 [95% CI 0.67–1.27]; p = 0.62) [32]. Earlier Gernsheimer et al. published a randomized, double-blind clinical trial with assessment of efficacy of TXA in reduction of bleeding in patients undergoing treatment for hematologic malignancies (chemotherapy or HSCT) when platelet counts was below 30 000/µL [33]. The primary endpoint defined as World Health Organization (WHO) grade ≥ 2 bleeding was observed with similar frequency 50.3% vs. 54.2% of patients receiving TXA or placebo (an adjusted OR = 0.83 (95%CI: 0.50–1.34; p = 0.44). There are no data on the use of above medications in cancer patients experiencing both VTE and thrombocytopenia.

TPO-RAs (romiplostim, eltrombopag, lusutrombopag, avatrombopag, hetrombopag) promote megakaryocyte growth, differentiation, and platelet production [34]. They have been approved for management of immune thrombocytopenia, aplastic anemia, hepatitis C-associated thrombocytopenia, and periprocedural thrombocytopenia in patients with chronic liver disease [35, 36].

The efficacy of TPO-RAs may be reduced in patients receiving chemotherapy due to cytotoxic and myelosuppressive effects of chemotherapy for platelets. NCCN guidelines recommend for chemotherapy induced thrombocytopenia (CIT) a consideration of platelet transfusions and chemotherapy dose adjustment/delay as well as additionally either (1) enrollment in a clinical trial of TPO-RAs or (2) off-label treatment with romiplostim but after a full discussion of benefits and harms [37]. There are some evidences on the long-term efficacy and safety of romiplostim for treatment of CIT [38]. Nevertheless, data on the effectiveness of other TPO-RAS in this indication is also available [39]. However there is a lack of high quality evidence (phase 3 trials) confirming efficient efficacy for TPO-RA in CIT [40]. None of the TPO-RAs have been approved by the FDA or the EMA for the treatment of CIT.

The International Society on Thrombosis and Haemostasis (ISTH) Subcommittee on Hemostasis and Malignancy summarized the current knowledge and experience on the use of TPO-RAs for CIT in various oncological scenarios [41]. According to the document if inadequate platelet recovery at day 1 of a chemotherapy cycle is observed, TPO-RAs may be considered to avoid chemotherapy dose reduction or a delay of ≥ 7 days. The suggested potential use of a TPO-RA (romiplostim over other TPORAs) may be in patients with solid tumors where full relative dose intensity chemotherapy may support beneficial prognosis. That is why TPO-RA should be continued for whole the duration of chemotherapy. Guidance statement recommends against the use of prophylactic tranexamic acid for the prevention of hemorrhage in CIT as well as against the use of TPO-RA for the management of CIT during therapy for acute myeloid leukemia, high-risk myelodysplasia, lymphoma. The document emphasizes that the available research results do not justify the use of TPO-RA for the management of CIT in hematopoietic stem cell transplantation (HSCT). Extensive data analyses do not confirm the efficacy of TPO-RA in reducing bleeding in hematological malignancies undergoing intensive chemotherapy or stem cell transplantation [42].

There is no clear information in the above documents regarding the potential use of TPO-RAs as an adjunctive therapy in cancer patients with VTE and grade 3 or 4 thrombocytopenia.

Moreover, VTE may be a theoretical complication of TPO-RAs, not clearly confirmed in clinical studies [43]. The thrombotic risk is more likely related to the underlying malignancy. However, theoretically, in the case of recurrent VTE while on therapeutic anticoagulation, discontinuation of romiplostim may be considered [44].

Rationale for anticoagulation in thrombocytopenia

Cancer itself predisposes to arterial and venous thromboembolism (ATE and VTE) [45–47]. Austrian population-level data estimated a prevalence of ATE and VTE 5.4% and 4.6% respectively in people with cancer [48]. Some histological and molecular diagnoses significantly increase this risk of VTE (up to a level of about 10–20%) [49–51]. Anticancer drugs can further increase this risk of VTE above 20% [52–55]. So we have the first indications for antithrombotic therapy. Older age is a classic risk factor for cancer and cardiovascular diseases, which often require antithrombotic therapy, either antiplatelet drugs or anticoagulation, depending on the diagnosis [56].

Thrombocytopenia in cancer patients does not reduce the risk of VTE [57]. Aspirin improved survival in cancer patients with acute myocardial infarction coexisting with thrombocytopenia [58, 59]. It confirms that prothrombotic readiness in cancer diseases is not only related to platelet function. Cancer-associated hypercoagulability is very complicated disorders. Cancer cells release chemotactic, angiogenic and pro-inflammatory mediators that activate circulating monocytes, neutrophils and endothelial cells. Other mechanism includes activated oncogenes (K-ras, EGFR, PML-RARalpha, and MET) or inactivated tumor suppressors (e.g., 53 or PTEN) which may induce the expression of tissue factor, a potent procoagulant molecule, and plasminogen activator inhibitor-1, a fibrinolysis inhibitor. Regardless of number, platelets can interact with tumor cells through different receptors [60, 61].

Cancer patients may be at increased risk of bleeding due to endothelial dysfunction and cancer-related hemostasis disorders [19]. If thrombocytopenia is a part of the history of a cancer disease development, or occurs as an adverse event of anticancer treatment, it usually complicates the possibility of continuing antithrombotic therapy. There are ongoing discussions among experts on how the platelet count limit should be set for the need to modify the method of anticoagulation [62].

Decisions based on severity of VTE

During anticancer treatment, when its continuation determines the patient’s survival, one of the solutions for primary and secondary prevention of bleeding related to thrombocytopenia may be platelet transfusions. This is particularly important when anticoagulant therapy is also necessary and grade 3 or 4 thrombocytopenia is present. Predicting the risk of thrombocytopenia associated with cancer drugs may also be useful for planning the most appropriate antithrombotic therapy from the beginning. The key seems to be the decision right away whether to choose low-molecular-weight heparins or oral anticoagulants. If the grade 3 or 4 thrombocytopenia is expected it seems to be justify dynamic monitoring of the patients’ morphology to modify anticoagulation at the appropriate moment.

Many additional factors may impact on choosing anticoagulation in thrombocytopenia. Complete suspension may be considered if the time from VTE index is longer than 30 days and the expected duration of grade 4 thrombocytopenia is long, whether it will be months or years. In case of expected grade 3 thrombocytopenia it is necessary to define additional factors influencing a risk of bleeding: history of bleeding, liver function, kidney function, prothrombin time [63]. The stage of the cancer, previous anticancer therapies, performance status may also elevate a risk of bleeding [64]. It is worth paying attention to the frailty syndrome [65, 66]. These factors will influence whether the primary prophylactic dose or half the therapeutic dose of low molecular weight heparin is selected.

The last but key issue is an indication for anticoagulation, more precisely, the expected benefit of anticoagulation compared to the life-threatening risks associated with its omission [67]. Benefit-harm ratio will be different for the acute phase of VTE treatment [68] compared to chronic/prolonged anticoagulation [69], and finally completely different in atrial fibrillation prophylaxis against stroke [70]. A separate problem is anticoagulation in patients with mechanical valves.

In VTE, the risk of recurrence or progression during anticoagulation (mainly in cohorts without thrombocytopenia) ranges from 10% to 15% for proximal lower limb thrombosis and segmental pulmonary embolism to 4% to 8% for thrombotic events considered low-risk, such as catheter-related thrombosis, distal lower limb thrombosis, or subsegmental pulmonary embolism [54]. The first month of anticoagulant treatment for VTE and especially pulmonary embolism is a high-risk period both in terms of the risk of recurrent bleeding and thrombosis [71]. Of course, this risk is even more increased in tumors with an increased risk of thrombotic events [72]. The highest incidence rates of VTE can be observed in pancreas, brain, liver oesophagus, lung, stomach, ovary cancer, as well as multiple myeloma and leukaemia [73].

In patients with grade 3 or 4 thrombocytopenia (platelets count < 50 000/µl) who require anticoagulation for acute thrombosis, two treatment strategies are often considered: (1) full-dose anticoagulation with higher target support via platelet transfusions (2) modified-dose anticoagulation (half the therapeutic dose or a prophylactic dose of the anticoagulant) [74].

There are no randomized trial data to support these strategies. Furthermore, the rationale for using modified-dose low molecular weight heparin (LMWH) in acute cancer related thrombosis (CAT) with thrombocytopenia comes from observations in the treatment of deep vein thrombosis (DVT) and not pulmonary embolism (PE). Nevertheless, the validity of this strategy has been confirmed in everyday practice [75].

Decisions based on characteristics of cancer disease

The problem of anticoagulation and thrombocytopenia should be discussed separately in three main groups of patients: (1) with solid tumors, (2) with hematological malignancies, (3) underwent hematopoietic stem cell transplantation. The risks of bleeding and thrombosis, the duration of thrombocytopenia and treatment options are different.

(1) Patients with solid tumors.

One in five patients with solid tumors and VTE has an inferior vena cava (IVC) filter implanted, although the indications are still a matter of debate [76]. The Polish study performed mainly in solid tumors confirmed that IVC filter placement is a safe and effective method of preventing PE in cancer patients with contraindications to anticoagulation [77]. Thrombocytopenia may be included in such understanding of the barriers of optimal anticoagulation [78]. IVC implantation seems to be useful in cancer patients with acute VTE and thrombocytopenia < 50 000/µl, especially if platelet transfusion is not considered beneficial. Furthermore, considering thrombocytopenia as a transient risk factor for bleeding, the filter can be removed if the platelet count improves and will be above 50 000/µl [79].

The benefit of IVC can be expected in cancer patients with acute lower extremity DVT and bleeding risk. Based on the available experiences from daily practice, the survival without pulmonary embolism development was significantly better among patients after placement of an IVC filter [80]. The benefit was clearly confirmed in tumor with all scales of thrombosis risk: very high (e.g., pancreaticobiliary cancer), high (e.g., lung), and low (e.g., prostate). The additional analysis of risk factors revealed the benefit was achieved without increase the risk of new DVT development. The subgroup with hematological malignancies also experienced significant benefits.

A retrospective observation was focused on patients with cancer-associated pulmonary embolism (PE) and a lower rate of anticoagulation use [81]. The benefit of filter implantation was expressed as a similar risk of recurrent PE. The patients with filter had poorer overall survival, which may reflect more advanced cancer disease and more strong contraindication to anticoagulation.

(2) Patients with hematological malignancies.

VTE in hematological malignancies affects mortality and its frequency is comparable to that in high-risk solid tumors [82]. The aggressiveness of the hemato-oncology disease determines the risk of VTE [83]. Patient-related risk factors, type of anticancer therapy, platelet and leukocyte count and the presence of infections are of additional importance [84].

The usefulness of LMWH in frail thrombocytopenic patients with hematological malignancies was assessed in two observational studies [57, 85]. Both were promising because excluded the significantly increased risk of bleeding. In patients with acute leukemia and non-CVC related VTE and severe thrombocytopenia (PLT < 30.000/µL), treatment with LMWH was effective if a full dose was used for one month and adjusted regimens for the following 3 months [86].

Frail subjects in hemato-oncology are patients with VTE under active chemotherapy and expected or overt thrombocytopenia with PLT < 50.000/µL [87]. Experts prefer to use LMWH and agree to stop anticoagulation when PLT < 20.000/µL, only in myelofibrosis and amyloidosis AL IVC filter is more appropriate option than only discontinuation. When PLT are between 20 and 50 000/µL, full therapeutic dose of LMWH shoul be continued in myeloproliferative and limphoproliferative including acute lymphoid leukemia, Hodgkin lymphoma, multiple myeloma. Only in patients with Non Hodgkin Lymphoma and PLT in range of 20 and 50 000/µL, it is justify to reduce by 50% the therapeutic dosage of LMWH.

The full-dose anticoagulation strategy especially in acute VTE should be supported by platelet transfusion to achieve a higher target platelet count (i.e. 40–50 000/µl) [88]. The CAVEaT (the Cancer Associated Venous Thrombosis and Thrombocytopenia) was a prospective multicenter cohort study [89] which included 105 patients with hematological malignancies, VTE diagnosed within 28 days, and platelets count < 50 000/µl. The primary objectives were to describe practice in the United Kingdom in relation to recommendations in national guidelines. The secondary objective was to check the development of new research. Anticoagulation was administered in 53% of patients with full intensity (LMWH or UFH), 28% with 50% dose LMWH, 3% with prophylactic dose LMWH, 4% with DOACs, and 11% with no anticoagulation. The study confirmed the heterogeneity of approaches and only 43% patients were managed using the guideline approach. Authors observed the high rates of bleeding: 7% major and 25% clinically relevant. There were 7% of thrombosis recurrence and 5% of major bleeding in group of patients on full dose LMWH/UFH. There were 13% of thrombosis progression and 13% of major bleeding in patients on reduced dose LMWH. The study revealed neither beneficial platelet transfusion threshold nor anticoagulant dose reduction threshold with clear reduction of the risk of thrombosis progression or major bleeding. Patients had a change of treatment strategy: 27% during 28 days, 51% during 90 days.

The TROVE (Thrombocytopenia Related Outcomes with Venous thromboEmbolism) was prospective observational study [90]. The preferred option in 62% of patients was full-dose anticoagulation (mainly LMWH) with platelet transfusion. Only 27% patients were treated with a modified dose and 11% did not receive anticoagulation. The majority of patients receiving the modified dose had diagnosed hematological malignancies (94%). Comparing full-dose and modified-dose anticoagulation during 60 days of observation, the rate of major hemorrhage was 12.8% and 6.6%, VTE recurrence was 5.6% and 0%, respectively. The biggest challenge in interpreting the TROVE study is that the baseline characteristics of the patients differed between those on full anticoagulation and those on modified therapy. Therefore, no conclusions can be drawn regarding a significant reduction in the risk of bleeding or VTE recurrence. Modifying anticoagulation may be a valuable option to consider in thrombocytopenia, however, the choice should also be determined by the thrombotic risk and a clinical condition of each patient.

Retrospective studies in hemato-oncology indicated that anticoagulation is a superior strategy to no anticoagulation or IVC filter [67, 91]. The British Society for Haematology [92] recommends platelet transfusion to maintain a platelet count of >40–50 000/uL and full dose anticoagulation with low molecular weight heparin (LMWH) for the first 3 months from the date of the VTE. If it is not possible, reduced doses of LMWH are recommended.

(3) Patients underwent hematopoietic stem cell transplantation.

The reported incidence of VTE can be high (up to almost 15%) in hematopoietic stem cell transplantation (HSCT) recipients and can lead to increased mortality [93]. Prolonged thrombocytopenia in post-transplantation period may be a real problem against anticoagulation [94]. One study reported that 34% of VTE occurred at a platelet count less than 50 000/µl and 13% at a platelet count less than 20 000/µl [13]. Bleeding occurred in more than 15% patients (3.6% of fatal bleeding) and was associated mainly with anticoagulation.

The key questions are whether to continue or temporarily discontinue anticoagulation and at what platelet count there is a real risk of bleeding. Lymphoma and myeloma patients with index VTE prior to HSCT were observed based on the severity of thrombocytopenia [95]. The study confirmed that temporarily discontinuation of anticoagulation during thrombocytopenia may be the best risk-benefit clinical decision, because there was no increased risk of VTE recurrence or bleeding, platelet transfusions were less frequently used. The study revealed hepatic and renal dysfunction (expressed by elevated bilirubin, creatinine or prothrombin time) were predictors of major bleeding, whereas platelet count had no significant predictive value. It seems worthwhile to develop a model for accurate prediction of VTE recurrence after HSCT in order to consciously discontinue anticoagulation, especially in patients who are not at risk [96, 97]. Moreover, it should be emphasized that discontinuation of anticoagulation is a valuable option for patients, but not those with acute VTE (less than 30 days since the onset of acute symptoms).

Should we prefer oral anticoagulation in hematological malignancies?

The HOKUSAI VTE Cancer study was the first published randomized, open-label, non-inferiority, phase III trial comparing dalteparin with edoxaban in treatment of VTE in cancer patients [98]. Patell et al. performed a post hoc analysis to compare outcomes in patients with platelet count ≤ 100 000/µl and patients without thrombocytopenia [99]. Among 1045 patients with mostly solid malignancies (89%) there was a subgroup of 101 (9.6%) patients with platelet counts ≤ 100 000/µl (at baseline, 1 month, or 3 months). They had a significantly higher risk of major bleeding (9.0% vs. 4.0%, SHR = 2.4; p = 0.02) and clinically significant bleeding (17.9% vs. 9.6%, SHR = 2.0; p = 0.01), but did not have a significantly higher risk of VTE recurrence (9.8% vs. 7.4%, SHR = 1.3; p = 0.37) or increased mortality (21.8% vs. 26.0%, HR = 0.9; p = 0.48). It may be concluded that even mild thrombocytopenia doubles the risk of major or clinically significant bleeding if cancer patients receive anticoagulation. The study also shows an interesting relationship regarding the choice of anticoagulation. Edoxaban, compared with dalteparin, caused more major bleeding in gastrointestinal cancers and thrombocytopenia (16.8% vs. 0%; p < 0.01). Dalteparin, compared with edoxaban, caused more bleeding if thrombocytopenia coexisted with hematological malignancies (19.0% vs. 0%; p < 0.01). There is a potential proof that LMWH may provoke more bleeding episodes in hematologic malignancies with thrombocytopenia. This requires further observation.

A retrospective analysis of cancer patients with VTE and thrombocytopenia (platelet count < 100 000/µl) showed that VTE recurrence and major bleeding within 90 days were similar in patients treated with LMWH and DOACs (direct oral anticoagulant inhibitors) [100]. However, the results of this study should be interpreted very carefully, because the number of patients was very small and the analyzed population was very diverse. The overall VTE recurrence rate was 2.4%, bleeding occurred in 19.0% patients (mainly clinically relevant non major bleeding). Unfortunately, the study was not randomized, more patients with hematological malignancies (acute leukemia, myelodysplastic syndrome, aggressive lymphoma) and grade 3 or 4 thrombocytopenia received LMWH and experienced an increased risk of bleeding. In such scenario, the real risk associated with DOACs may be underestimated. However, there are data showing that patients with hematological malignancies, despite the more frequent co-existing grade 3 or 4 thrombocytopenia, experience bleeding and VTE recurrences less frequently [101].

Before considering DOACs in cancer patients with VTE and thrombocytopenia, we need more data to determine whether this is a safe and patient-protecting decision.

Outcome defined as bleeding occurrence versus VTE recurrence

In a meta-analysis of 707 patients, mainly with hematological malignancies, with thrombosis and thrombocytopenia < 100 000/µl, three strategies were compared: full-dose, modified-dose, and no anticoagulation [102]. The main anticoagulant drug used was LMWH. The VTE recurrence rate was high: 2.65/100 patient-months on full dose and 3.51/100 patient-months on modified dose, and 3.68/100 patient-months without anticoagulation. Major bleeding was reported even more frequently at 4.45/100 patient-months with the full dose and 4.16/100 patient-months with the modified dose, and 2.2/100 patient-months if anticoagulation was not used. The incidence of all bleedings was very high: 13.20/100 patient-months for full anticoagulation, 10.16/100 patient-months for modified anticoagulation and the lowest: 3.89/100 patient-months if there was no anticoagulation.

Houghton et al. [67] showed that sometimes omitting anticoagulation is very beneficial for reducing the risk of bleeding if platelet counts < 50 000/µl in hematologic malignancies. Bleeding occurred in 27% and 3%, respectively (almost a tenfold difference), whereas recurrent VTE occurred in 2% and 15% (almost a sevenfold difference) of patients receiving or not receiving anticoagulation. In the discussed study anticoagulation meant prophylactic, intermediate-dose, or therapeutic dose. Importantly from a practical point of view, bleedings occurred mainly early in the first 30 days, and VTE recurrences mainly after the 40th day of observation. Such situation may guarantee a chance to recover from grade 3 or 4 thrombocytopenia if it is temporarily induced by chemotherapy.

Moreover, it sometimes happens, that the difference in effectiveness of modified, i.e. reduced, anticoagulation is not significantly better than no anticoagulation. Kopolovic et al. [103] showed in the population of different cancer diseases there was a significant difference in the rate of bleeding between patients on reduced dose of anticoagulation and patients with interrupted anticoagulation: 33% vs. 12%. The rate of VTE recurrence was similar 44% vs. 47%. Therefore, the most important thing is an individual approach to each case (Table 3).

Table 3.

Observations with benefit of interruption of anticoagulation in comparison to continuation of anticoagulation in patients with cancer disease, VTE and thrombocytopenia ≤ 50,000/µL

First Author	Bleeding occurrence		VTE recurrence
	Continuation of anticoagulation	Interruption of anticoagulation	Continuation of anticoagulation	Interruption of anticoagulation
Houghton et al. [67]	27%	3%	2%	15%
Kopolovic et al. [103]	33%	12%.	44%	47%

On the other hand, observational studies are available that clearly support the benefit of adopting a strategy with modified dose of anticoagulation [57, 75] (Table 4). The benefit is based on a reduction in the risk of bleeding while maintaining effectiveness. However, not all studies confirm this benefit so unequivocally. In many cases, personalization of the anticoagulation choice will be decisive, taking into account not only the platelet count, but many other factors determining bleeding risk, those related to the patient’s history, the characteristics of the cancer disease and the type of anticancer treatment.

Table 4.

Observations with comparison between full and reduced anticoagulation in patients with cancer disease, VTE and thrombocytopenia ≤ 50,000/µL

First Author	Bleeding occurrence		VTE recurrence
	Full dose	Modified dose	Full dose	Modified dose
Khanal et al. [57]	26.7%	4.5%	33.3%	13.6%
Mantha et al. [75]	9.8%	4.2%	20.6%	20.8%

Recommendations of scientific societies

In recent years, several documents of guidelines for the treatment of cancer related VTE have appeared in medical literature. American Society of Hematology (ASH) published last guidelines in 2021 [104], European Society for Medical Oncology (ESMO) announced guidelines in electronic version in January 2023 [105], American Society of Clinical Oncology (ASCO) presented online version in April 2023 [106]. The last important document was published in 2024 by the National Comprehensive Cancer Network (NCCN) [107].

Importantly, there are also two scientific documents dedicated directly to cancer patients with VTE and thrombocytopenia published in 2018 by the International Society on Thrombosis and Haemostasis (ISTH) Scientific and Standardization Committee (SSC) on Haemostasis and Malignancy [108] and the second document published in 2022 by the European Hematology Association (EHA) in cooperation with the European Society of Cardiology (ESC) [109].

There is evidence that optimal anticoagulation does not induce increased bleeding if the platelet count is above 50 000/µl [57]. It is obvious that there must be no additional risk factor for bleeding (liver disorders, renal dysfunction, etc.). Therefore, all documents agree not to modify anticoagulation until platelets count is >50 000/µl.

ASH document reminds thrombocytopenia in the context of heparin-induced thrombocytopenia (HIT) and proposes that fondaparinux might be considered for patients with cancer and VTE and a history of HIT similarly to the previous document [110].

The last ASCO guidelines document does not focus on the problem of thrombocytopenia. The previous ASCO document listed thrombocytopenia as a barrier to therapeutic anticoagulant therapy in patients with cancer: either an absolute (persistent thrombocytopenia < 20 000/µl) or relative contraindication (persistent thrombocytopenia < 50 000/µl) [111]. This document also did not present any recommendations for cancer patients with VTE and thrombocytopenia.

Three societies have decided to define subgroups of cancer patients requiring special consideration when selecting anticoagulation for VTE coexisting with thrombocytopenia when platelets count is below 50 000/µl (Table 5). ESMO defined risk of thrombosis recurrence or extension [72]. ISHT and ESMO prefer anticoagulation in acute VTE understood as the first 30 days from diagnosis [112], because after this time the risk of VTE recurrence decreases significantly [113]. It is worth to highlight the three main documents understand consideration of using anticoagulation (full or reduced dose) instead of giving up (Table 6).

Table 5.

Definitions proposed by scientific societies to identify of specific subgroups of VTE cancer patients with platelets count < 50 000/µl who can benefit from using anticoagulation instead of interruption

ISHT	EHA	ESMO
- Acute VTE, i.e. within the first 30 days of diagnosis - High-risk VTE: Pulmonary embolism in segmental or more proximal vessels, Proximal deep vein thrombosis, Recurrent/progressive thrombosis.	- Very high risk: pulmonary embolism with hemodynamic instability - High risk: pulmonary embolism within 6 weeks of diagnosis or acute DVT	- High risk of thrombus progression: acute VTE (i.e. first 30 days from diagnosis), segmental or more proximal PE, proximal DVT or a history of recurrent thrombosis, - Low risk of thrombus progression: >30 days from diagnosis, distal DVT, isolated subsegmental PE.

Table 6.

The summary of the most important recommendations focused on VTE treatment in thrombocytopenic cancer patients

Authorship	Recommendations	Level of evidence
ISHT	• In patients with high risk acute VTE (< 30 days) and platelets count < 50 000/µl, full-dose anticoagulation (LMWH/UFH) and platelet transfusion to maintain platelet count ≥ 40–50 000/µl • In patients with lower risk acute VTE or VTE on chronic treatment (> 30 days): Reduce the LMWH dose to 50% of the therapeutic dose or use a prophylactic LMWH dose if platelets count is 25–50 000/µl Temporarily discontinue anticoagulation if platelets count is < 25 000/µl	Moderate consensus among the panel members
EHA	• In patients at very high risk: therapeutic dose of LMWH and platelet transfusions with a target platelet count of 40–50 000/µl, if achievable, may be considered for a maximum of 14 days. • In patients with high-risk PE and in centers with appropriate expertise the following treatment options may be considered through individualized approach: systemic thrombolysis, interventional procedures of thrombus removal including catheter-based thrombolysis or pharmacomechanical catheter-directed reperfusion techniques • TPO-RA may be used in patients with anticipated long duration of thrombocytopenia, but not in patients with high-thrombotic risk, acute leukemia, MDS, or extensive bone marrow infiltration.	Expert opinion
	• If platelets count is in the range of 25–50 000/µl: LMWH in prophylactic or therapeutic doses reduced by 50% should be used in patients with acute VTE • If platelets count is < 25 000/µl : stopping anticoagulation • Removable inferior vena cava filter (IVCF) may be considered on an individual basis in patients with acute PE or acute lower extremity DVT up to 30 days since the diagnosis.	Individual cohort study including low-quality randomized control study
	• Removal of the central venous catheter (CVC) in patients who cannot receive anticoagulation, within the first 30 days of an acute upper extremity deep vein thrombosis (UE-DVT).	Case series and poor quality cohort and case-control studies
ESMO	• In high risk: full-dose anticoagulation may be considered in combination with platelet transfusion support aiming at platelet count > 40–50 G/L. • In low risk: intermediate- to prophylactic-dose LMWH may be considered with temporary discontinuation of anticoagulation if the platelet count falls below 25 G/L.	Not available

The 2024 NCCN Clinical Practice Guidelines in Oncology discuss chemotherapy-induced thrombocytopenia without any presentation of detailed definition and only in terms of contraindications to therapeutic anticoagulation and management of anticoagulation for VTE in a few scenarios (Table 7).

Table 7.

VTE treatment in cancer patients with thrombocytopenia according to the 2024 NCCN clinical practice guidelines in oncology

Cancer-Associated Venous Thromboembolic Disease	Proposed management in thrombocytopenia
Pulmonary embolism	→ Consider filter placement if unable to treat with anticoagulation within 1 month of onset of symptomatic pulmonary embolism [114] (retrievable filter preferred) ± Consider embolectomy for treatment of massive pulmonary embolism (category 2B) → Follow frequently for change in clinical status → Recommend IVC filter removal, if tolerating anticoagulation therapy
Proximal lower extremity Pelvic/iliac/inferior vena cava (IVC) or femoral/popliteal	→ IVC filter (retrievable filter preferred)
Distal lower extremity Peroneal, anterior and posterior tibial, and muscular (soleus and gastrocnemius)	→ Follow-up with serial US → If progression to proximal vein → IVC filter (retrievable filter preferred)
Upper limb/chest Brachiocephalic, subclavian, axillary, internal jugular, brachial or superior vena cava (SVC)	→ Follow until contraindication is resolved or progression of DVT → Re-evaluate for risk/benefit of anticoagulation → Elements for consideration in decision “Not to Treat”
Catheter-related DVT	→ Remove catheter or follow with serial imaging → Re-evaluate for risk/benefit of anticoagulation → Elements for consideration in decision “Not to Treat”
Acute Hepatic Vein Thrombosis (symptoms/signs ≤ 8 weeks)	→ Hepatology evaluation → Consider transjugular intrahepatic portosystemic shunt (TIPS) or surgical shunt
Acute Portal, Mesenteric, and/or Splenic Vein Thrombosis (symptoms/signs ≤ 8 weeks)	→ GI/surgery evaluation → Surgery (if bowel infarction)

Conclusions

In the absence of high-quality clinical trials, clinicians treating VTE in cancer patients with thrombocytopenia are required to make an individual decision for each patient. Such decision is determined by the severity and duration of thrombocytopenia, which largely depends on whether the patient is undergoing anticancer treatment for a solid tumor or a hematological malignancy. Patients after HSCT deserve special attention. There are three therapeutic options: full dose of anticoagulation with platelet transfusions, reduced dose or temporary suspension of anticoagulation. The most important is to use an appropriate option for a selected patient.

Author contributions

All authors wrote the main manuscript text and prepared tables.All authors reviewed the manuscript.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Competing interests

Sebastian Szmit received speaker fee: Novartis, BMS, Bayer, Pfizer, Astra Zeneca. Other authors declare no competing interests.