Anticoagulation in individuals with thrombocytopenia

INTRODUCTION — Thrombocytopenia may increase bleeding risk, but it does not protect against venous thromboembolism (VTE) or stroke. Thus, caring for patients with both thrombocytopenia and an indication for anticoagulation (eg, VTE prophylaxis or treatment, stroke prophylaxis or treatment) can be challenging. Evidence to guide appropriate therapy in this setting is very limited.

This topic discusses our approach to the use of anticoagulation in an individual with thrombocytopenia, including decisions about the need for anticoagulation, anticoagulant dosing, therapies to raise the platelet count, and alternatives to anticoagulation if the bleeding risk is thought to be too high.

Related subjects are discussed in separate topic reviews:

●Cancer

•Risk and prevention of VTE – (See "Risk and prevention of venous thromboembolism in adults with cancer".)

•VTE treatment – (See "Anticoagulation therapy for venous thromboembolism (lower extremity venous thrombosis and pulmonary embolism) in adult patients with malignancy".)

●Stroke

•Risk in patients with atrial fibrillation – (See "Atrial fibrillation in adults: Use of oral anticoagulants".)

•Treatment in atrial fibrillation – (See "Stroke in patients with atrial fibrillation".)

●Thrombocytopenia

•Causes – (See "Diagnostic approach to thrombocytopenia in adults" and "Drug-induced immune thrombocytopenia".)

•Liver disease – (See "Hemostatic abnormalities in patients with liver disease".)

GENERAL PRINCIPLES OF TREATMENT

Overview of decision-making — Decisions about whether to use anticoagulation in patients with thrombocytopenia must balance the risks of thrombosis and thrombosis progression with the risks of bleeding due to thrombocytopenia as well as due to other bleeding risk factors. Estimates of bleeding and thrombosis incidence in thrombocytopenic patients have been reported in some settings [1-4]. However, these data have limitations and may not be available for specific scenarios.

●There is no simple formula for calculating which risk (thrombosis or bleeding) is greater

●There is no anticoagulant that can reduce thrombotic risk without also increasing bleeding risk

The expected duration of these risks may also factor into decision-making, along with the patient's values and preferences around which risks they feel most strongly about avoiding. The treatment plan should be reassessed based upon the actual and anticipated trajectory of the platelet counts, since thrombocytopenia may be dynamic, especially in the context of myelosuppressive cancer treatment.

Once the decision to use anticoagulation has been made, the choice of anticoagulant is based on the underlying indication and other patient factors. The choice of anticoagulant and risk of bleeding with different anticoagulants is discussed separately in topics on specific indications.

General information is presented in the following sections. Guidelines on venous thromboembolism and atrial fibrillation in the setting of thrombocytopenia are available from the International Society on Haemostasis and Thrombosis and the European Hematology Association [5,6].

Specific clinical scenarios such as cancer-associated venous thromboembolism (VTE) and other common combinations of thrombocytopenia and thrombosis risk are presented below. (See 'Cancer-associated VTE' below and 'Other selected clinical scenarios' below.)

Estimating and managing bleeding risk — The presence of one or more bleeding risk factors increases concerns about bleeding that may be exacerbated by the use of an anticoagulant. Some risk factors are known from the general population, while others have been assessed in patients with thrombocytopenia with or without anticoagulation [5]. Commonly cited risk factors include the following [5,7,8]:

●Recent major bleeding – Bleeding risk is probably greatest in individuals who have already had a clinically serious bleed, especially if the bleeding occurred at a platelet count >50,000/microL and if the source of bleeding has not been treated.

●Platelet count <50,000/microL – It is generally accepted that bleeding risk is higher in individuals with platelet counts <50,000/microL, and that severe spontaneous bleeding is most likely with platelet counts <10,000/microL, especially if due to an underlying bone marrow disorder rather than due to immune destruction of platelets. However, there is not a good linear correlation between the platelet count and the risk of serious bleeding in thrombocytopenic individuals. Studies have shown that the bleeding risk increases even with mild thrombocytopenia receiving anticoagulation for VTE or atrial fibrillation (<100,000/microL) [3,4]. Whether this finding reflects an actual bleeding risk with platelet counts of 50,000 to 100,000/microL, an underlying disease state, or a marker of subsequent (possibly undocumented) platelet counts <50,000/microL remains to be determined. (See "Diagnostic approach to thrombocytopenia in adults", section on 'When to worry about bleeding'.)

●Hematopoietic stem cell transplantation – Individuals who have undergone hematopoietic stem cell transplantation, especially allogeneic, are at increased risk of bleeding. This is likely due to several factors including decreased platelet production, toxicities of the transplant conditioning regimen, and graft-versus-host disease.

●Cancer – Active cancer is associated with an increased risk of bleeding. Acute leukemia, in particular acute promyelocytic leukemia, is associated with an especially high bleeding risk. Primary and metastatic brain cancer are associated with an increased risk of intracranial bleeding relative to cancer patients without brain involvement. (See "Clinical manifestations, pathologic features, and diagnosis of acute promyelocytic leukemia in adults", section on 'Coagulopathy and APL' and "Overview of the clinical features and diagnosis of brain tumors in adults", section on 'Clinical manifestations'.)

●Coagulation abnormalities – Coagulation abnormalities may be acquired due to conditions such as severe liver disease or disseminated intravascular coagulation. Inherited deficiencies in coagulation factors are more uncommon, but clinicians should be aware of this possibility, which may affect bleeding risk. (See "Hemostatic abnormalities in patients with liver disease" and "Evaluation and management of disseminated intravascular coagulation (DIC) in adults" and "Approach to the adult with a suspected bleeding disorder", section on 'Family history'.)

●Platelet function disorders – Conditions that interfere with platelet function (eg, uremia from kidney failure) or certain medications (eg, tyrosine kinase inhibitors such as ibrutinib) may further increase bleeding risk independent of platelet number. (See "Uremic platelet dysfunction" and "Non-cardiovascular toxicities of molecularly targeted antiangiogenic agents", section on 'Bleeding'.)

Inherited platelet function disorders are much less common but may be relevant for some individuals. (See "Inherited platelet function disorders (IPFDs)".)

●Chronic kidney disease – Chronic kidney disease stage III or greater is associated with an increased bleeding risk. The reason for this is multifactorial, including effects on excretion of anticoagulants and effects on platelet function [9].

●Older age – Older age (>60 years), especially >75 years, is associated with increased bleeding risk.

●Fall risk – Some individuals may be at increased risk of bleeding due to conditions that increase their risk of falls or other injuries. This is most relevant for outpatients receiving anticoagulation; during hospitalization, measures are commonly taken to minimize this risk. (See "Falls in older persons: Risk factors and patient evaluation" and "Falls: Prevention in community-dwelling older persons".)

For individuals with VTE or atrial fibrillation who do not have one of these increased bleeding risk factors, anticoagulation can generally be given if the platelet count is ≥50,000/microL. If the platelet count is <50,000/microL, decision-making is individualized and takes into account the likelihood of VTE progression or stroke, the anticipated duration and trajectory of thrombocytopenia, and the available options for increasing the platelet count, which depend on the reason for the thrombocytopenia, as discussed below and in related topic reviews. (See 'Cancer-associated VTE' below and "Early antithrombotic treatment of acute ischemic stroke and transient ischemic attack" and 'Platelet count support' below.)

For individuals with VTE who have an increased risk of bleeding, there are a number of options for preventing VTE progression or recurrence. This includes the use of lower-dose anticoagulation (either prophylactic dose or reducing the dose by one-half), temporarily holding the anticoagulant (especially if bleeding risk factors are transient), placement of an inferior vena cava (IVC) filter (primarily when anticoagulation is held in the setting of acute pulmonary embolism [PE] or proximal lower extremity deep vein thrombosis [DVT]), and removal of the central venous catheter (in case of symptomatic catheter related VTE, when anticoagulation is held).

In general, addition of antiplatelet agents to anticoagulation is avoided when platelet counts are <50,000/microL, unless there is a very strong indication such as an intracoronary stent in the past 12 months, in which case anticoagulation dose reduction may be considered. Decisions of which of these options to use are highly patient-dependent.

●For VTE treatment, the importance of using anticoagulation is most influenced by the acuity of an existing thrombosis, as discussed below. (See 'Estimating risk of VTE and/or VTE progression' below.)

●For stroke prevention, decision-making is individualized according to several risk factors, as discussed in separate topic reviews. It is worth noting that cancer patients, who have a high prevalence of thrombocytopenia, also have an increased stroke risk relative to patients without cancer if they have atrial fibrillation [10]. (See "Overview of secondary prevention of ischemic stroke" and "Atrial fibrillation in adults: Use of oral anticoagulants".)

Estimating risk of VTE and/or VTE progression — The risk of VTE is greatest in the setting of a strongly prothrombotic risk factor, such as orthopedic or major abdominal surgery, certain cancers, active chemotherapy, and certain particularly prothrombotic disorders such as antiphospholipid syndrome, paroxysmal nocturnal hemoglobinuria (PNH) (unless receiving anti-complement therapy), or heparin-induced thrombocytopenia. (See 'Antiphospholipid syndrome' below and 'PNH' below and 'Heparin-induced thrombocytopenia' below.)

The risk of VTE progression (or recurrence) is greatest in the initial 30 days after the acute VTE event. Thus, after 30 days, thrombotic risk is expected to decrease, and these individuals are treated as low to intermediate thrombotic risk, depending on other risk factors. Other risk factors for VTE progression are related to the thromboembolism size (eg, single subsegmental versus lobar pulmonary embolism) and location (eg, upper extremity versus low extremity deep vein thrombosis), and whether the risk factor(s) for VTE were transient (eg, surgery) or persistent (eg, metastatic cancer), or whether the VTE was unprovoked.

For individuals with cancer, the presence of active cancer or chemotherapy and lower performance status are strong risk factors for VTE progression.

We consider the following individuals at higher risk for VTE progression/recurrence/embolization:

●VTE within the prior 30 days

●History of recurrent VTE, especially if recurrence followed short periods of anticoagulation interruption

●Proximal lower extremity DVT

●Segmental or larger PE

●Poor performance status (eg, bedridden for most or all of the day)

●IVC filter in place without anticoagulation increases the likelihood of DVT progression

●Active cancer and/or receiving chemotherapy (see 'Cancer-associated VTE' below)

●Strongly prothrombotic syndrome such as antiphospholipid syndrome, PNH, or heparin-induced thrombocytopenia (see 'Antiphospholipid syndrome' below and 'PNH' below and 'Heparin-induced thrombocytopenia' below)

●High-risk thrombophilia such as homozygous factor V Leiden or antithrombin deficiency

However, we do not advocate routine testing for these thrombophilias. Indications for testing are presented separately. (See "Evaluating adult patients with established venous thromboembolism for acquired and inherited risk factors".)

We consider the following individuals at lower to intermediate risk for VTE progression/recurrence/embolization:

●VTE that occurred more than 30 days prior

●Isolated distal DVT

●Isolated subsegmental PE (with negative bilateral lower extremity ultrasound)

Cancer patients with isolated distal DVT have a higher VTE recurrence risk than patients without cancer [11]. While the risk of recurrent VTE may be similar among cancer patients with proximal DVT and isolated distal DVT, the risk of fatal PE may be lower in patients with isolated distal DVT [11]. Similarly, there is evidence that selected patients without cancer with isolated subsegmental PE may be safely observed without anticoagulation [12-14].

●Higher risk for progression/recurrence – We generally use anticoagulation in patients with a prior VTE at higher risk for VTE progression or recurrence, especially if the VTE occurred in the prior 30 days (algorithm 1). The patient’s bleeding risk dictates whether we use full dose and platelet transfusion support (standard bleeding risk) or reduced dosing (higher bleeding risk).

●Lower or intermediate risk for progression/recurrence – For patients with a VTE and a lower to intermediate risk for VTE or VTE progression, we use anticoagulation if the platelet count is >50,000/microL and there are no other major bleeding risk factors. However, this decision is highly individualized according to the bleeding and thrombotic risks and the values and preferences of the individual patient.

Estimating risk of arterial thrombosis — The risk of arterial thrombosis is highly dependent on the patient's underlying condition, as discussed in separate topic reviews and in the sections below. (See 'Atrial fibrillation' below and 'Disorders that simultaneously cause thrombosis and thrombocytopenia' below.)

Similar to VTE risk, thrombocytopenia does not protect against arterial thrombosis in the general population and in specific pro-thrombotic settings such as cancer or immune thrombocytopenia. This is exemplified by a series of individuals with a number of underlying risk factors such as atrial fibrillation or antiphospholipid syndrome who had immune thrombocytopenia [15]. There were several instances in which discontinuation of anticoagulation for thrombocytopenia, even temporarily, was associated with thrombosis; in some cases, the thromboses were fatal.

Duration of anticoagulation — The duration of anticoagulation depends on whether the risk factors for bleeding and thrombosis were transient or persistent. Details that apply to different patient populations are discussed in separate topic reviews:

●Cancer

•General – (See "Anticoagulation therapy for venous thromboembolism (lower extremity venous thrombosis and pulmonary embolism) in adult patients with malignancy", section on 'Follow-up' and "Anticoagulation therapy for venous thromboembolism (lower extremity venous thrombosis and pulmonary embolism) in adult patients with malignancy", section on 'Duration of anticoagulation'.)

•Brain tumors – (See "Treatment and prevention of venous thromboembolism in patients with brain tumors".)

•Pancreatic cancer – (See "Supportive care for locally advanced or metastatic exocrine pancreatic cancer", section on 'Venous thromboembolism'.)

•Multiple myeloma – (See "Multiple myeloma: Prevention of venous thromboembolism".)

●Atrial fibrillation – (See "Atrial fibrillation in adults: Use of oral anticoagulants".)

●Liver disease – (See "Hemostatic abnormalities in patients with liver disease" and "Acute portal vein thrombosis in adults: Clinical manifestations, diagnosis, and management" and "Chronic portal vein thrombosis in adults: Clinical manifestations, diagnosis, and management".)

●Postsplenectomy – (See "Surgical management of splenic injury in the adult trauma patient" and "Elective (diagnostic or therapeutic) splenectomy", section on 'Venous thromboembolism'.)

●Antiphospholipid syndrome – (See "Management of antiphospholipid syndrome".)

●Paroxysmal nocturnal hemoglobinuria – (See "Treatment and prognosis of paroxysmal nocturnal hemoglobinuria", section on 'PNH with thrombosis'.)

Platelet count support — If a decision is made to use anticoagulation and thrombocytopenia is <50,000/microL, there are relatively few options to increase the platelet count.

Platelet transfusions — Platelet transfusions can be administered on an inpatient or outpatient basis to raise the platelet count and allow anticoagulation. When used to facilitate therapeutic anticoagulation, we favor a target platelet count of 50,000/microL. The frequency of platelet administration varies by patient and clinical scenario and is determined by monitoring the platelet count. This strategy is usually used for a period of days to weeks, until the acutely high thrombotic risk (or the thrombocytopenia) subsides.

Thrombopoietin receptor agonists (TPO-RAs) — Thrombopoietin receptor agonists (TPO-RAs; also called TPO-mimetics) are not optimal candidates to raise the platelet count to allow anticoagulation for acute VTE, since their effect is not immediate and adequate platelet response may take days to weeks.

In addition, there is some concern regarding increased risk of thrombosis with TPO-RAs; one of the TPO-RAs, eltrombopag, has been associated with an increased risk of portal vein thrombosis in patients with cirrhosis, while there are conflicting data regarding thrombotic risk with TPO-RAs in patients with immune thrombocytopenia, and this remains an area of active research and debate [16]. (See "Clinical applications of thrombopoietic growth factors", section on 'Lack of increased thrombosis in ITP'.)

The main consideration is whether there is evidence showing that TPO-RAs improve platelet counts in the specific setting leading to thrombocytopenia. TPO-RAs are approved for individuals with liver disease undergoing procedures and in patients with immune thrombocytopenia. There are some data showing that specific TPO-RA agents may improve platelet counts in patients with persistent chemotherapy-induced thrombocytopenia in order to enable continued administration of the culprit chemotherapeutic agent [17].

However, there is little published information on the usefulness of TPO-RAs to facilitate the administration of anticoagulation, and we do not routinely use these agents as a means of facilitating anticoagulation outside accepted indications such as immune thrombocytopenia. (See "Second-line and subsequent therapies for immune thrombocytopenia (ITP) in adults", section on 'TPO receptor agonists' and "Hemostatic abnormalities in patients with liver disease", section on 'General approach to invasive procedures'.)

Correction of the underlying disorder — In patients with immune thrombocytopenia and a strong indication for therapeutic-dose anticoagulation, it may be appropriate to start or change immune thrombocytopenia-directed therapy at a higher platelet count (eg, 50,000/microL) than would be used in patients without additional bleeding risk factors, for whom immune thrombocytopenia-specific therapy is generally initiated at a platelet count of 20,000 to 30,000/microL. (See "Initial treatment of immune thrombocytopenia (ITP) in adults", section on 'Whom to treat (indications for therapy)'.)

It is also worthwhile to confirm that the cause of the thrombocytopenia has been correctly attributed and there are no other interventions available that might improve the platelet count. There are numerous potential causes of thrombocytopenia including drug-induced immune thrombocytopenia, infection, and vitamin B12 deficiency that may be contributing to a low platelet count. In a series of 101 individuals with multiple myeloma undergoing hematopoietic stem cell collection, the incidence of heparin-induced thrombocytopenia was higher than expected at 4 percent [18]. Causes of thrombocytopenia are reviewed separately. (See "Diagnostic approach to thrombocytopenia in adults" and "Approach to the child with unexplained thrombocytopenia".)

CANCER-ASSOCIATED VTE

Risk factors for bleeding and thrombosis in cancer — Bleeding and thrombosis risks are both increased in individuals with cancer, and sometimes the same individual can have both bleeding and thrombosis (algorithm 1).

●Bleeding – The risk of bleeding in individuals with malignancy is higher than in the general population, and this extends to anticoagulation-associated bleeding. This is due to many factors, and the contribution of thrombocytopenia to bleeding risk is likely to be greatest when it is due to bone marrow suppression, as opposed to other causes such as hypersplenism or drug-induced immune platelet destruction. (See 'Estimating and managing bleeding risk' above.)

●Thrombosis – The risk of cancer-associated venous thrombosis is especially high with metastatic tumors, tumors causing venous compression or obstruction, certain tumor types (eg, pancreatic or gastric cancer, primary brain tumors, acute lymphoblastic leukemia, multiple myeloma) and certain cancer therapies (eg, asparaginase, immunomodulatory drugs [IMiDs; thalidomide, lenalidomide, pomalidomide]) [19]. The overall/cumulative risk of VTE in individuals with cancer has been estimated at approximately 10 to 15 percent, as discussed in detail separately. (See "Risk and prevention of venous thromboembolism in adults with cancer", section on 'Incidence and risk factors' and "Overview of the causes of venous thrombosis", section on 'Malignancy'.)

Anticoagulation is generally well tolerated, and mortality from VTE may be greater than mortality from bleeding in most cancer populations. As an example, in data derived from the RIETE registry (a large Spanish registry that included over 40,000 patients with VTE), the rate of fatal PE in patients with VTE and a platelet count <80,000/microL was 3.6 percent, compared with a rate of fatal hemorrhage of 2.0 percent [20]. Both of these rates were significantly higher than seen in participants with normal platelet counts. The benefit of anticoagulation in reducing thrombosis risk is presented below. (See 'Supporting evidence' below.)

Exceptions to the greater risk of mortality from thrombosis than bleeding include hematopoietic stem cell transplant recipients, for whom bleeding (including fatal bleeding) was more common than VTE; individuals with brain tumors who have an increased bleeding risk; and catheter-associated VTE, which may be treated locally (eg, with catheter removal). (See 'Special populations' below.)

Primary VTE prophylaxis — When indicated, primary VTE thromboprophylaxis is generally considered safe with platelet counts above 50,000/microL [21].

The approach to prophylactic anticoagulation for individuals with platelet counts <50,000/microL is more nuanced, and data are limited [22]. We generally do not use anticoagulation for VTE prophylaxis if the platelet count is below the 20,000 to 30,000/microL range. For individuals with platelet counts above this range but below 50,000/microL, use of prophylactic anticoagulation is individualized; it may be appropriate in those with an especially high thromboembolic risk (eg, high-risk malignancy, history of VTE). (See 'Estimating risk of VTE and/or VTE progression' above.)

VTE treatment/secondary prophylaxis

Platelet count 50,000/microL or above — In patients with platelet counts ≥50,000/microL who require anticoagulation for VTE treatment and/or secondary prophylaxis, full-dose anticoagulation is generally appropriate, as in nonthrombocytopenic populations. (See "Anticoagulation therapy for venous thromboembolism (lower extremity venous thrombosis and pulmonary embolism) in adult patients with malignancy".)

Nonetheless, cancer patients with VTE and platelet counts 50,000 to 100,000/microL also have an increased bleeding risk [3]. In these patients, close monitoring of the platelet count is warranted, especially if the nadir of chemotherapy-induced thrombocytopenia has not yet occurred.

There are certain exceptions, such as individuals who have increased bleeding risk factors:

●Recent major bleeding episode, especially if the bleeding source has not been treated

●Allogeneic hematopoietic stem cell transplantation

●Coagulation abnormalities (including due to liver disease)

●Platelet function disorders (including inherited and acquired disorders; the latter may be due to kidney or liver disease or certain medications)

●Age >75 years

●Increased risk for falls

Additional details about these risk factors are presented above. (See 'Estimating and managing bleeding risk' above.)

For individuals with mild-to-moderate thrombocytopenia (platelet count between 50,000 and 149,000/microL), VTE, and one or more of these increased bleeding risk factors, we generally still use full-dose anticoagulation, especially for proximal lower extremity DVT and segmental or more proximal PE.

Platelet count <50,000/microL — Our approach to VTE treatment in individuals with cancer who have a platelet count below 50,000/microL depends on three factors: bleeding risk, risk for VTE progression or recurrence, and in some cases, absolute platelet count.

Our approach to incorporating these factors is illustrated in the algorithm (algorithm 1). Key features include the following:

●For most individuals with cancer-associated VTE and platelet counts between 25,000 and 50,000/microL, we suggest anticoagulation. The lower threshold is an approximation, not an absolute cutoff. This practice is based on evidence from observational studies, as described below. (See 'Supporting evidence' below.)

●The dose of the anticoagulant (full therapeutic dose versus half-dose or prophylactic dose) is initially stratified by the presence or absence of bleeding risk factors other than thrombocytopenia, and the thrombotic risk.

In the absence of randomized trial data, there is true equipoise regarding optimal anticoagulation in patients with an acute VTE in the setting of concurrent thrombocytopenia. In a prospective, observational study of 121 patients with different cancer types, acute VTE and platelet counts <100,000/microL, 62 percent received full-dose anticoagulation and 27 percent received modified-dose anticoagulation [1]. At 60 days, the cumulative incidence of major hemorrhage was 12.8 percent in the group receiving full-dose anticoagulation and 6.6 percent in the group that received modified-dose anticoagulation. The cumulative incidence of recurrent VTE at 60 days was 5.6 percent in the full-dose and 0 percent in the modified-dose anticoagulation groups. A small number of patients with acute PE were included in the modified-dose cohort.

While these data generate the hypothesis that modified-dose anticoagulation may be safe and effective in this setting, a meta-analysis of observational data [23] and a subsequent prospective study [2] (patients with acute VTE, platelet counts <50,000/microL, and hematological malignancy) highlight the need for further research [2,23].

●Without strong bleeding risk factors – For individuals with standard bleeding risk (eg, lacking additional bleeding risk factors such as age >75 years, recent severe bleeding, hematopoietic stem cell transplantation, kidney or liver failure, or increased risk for falls in the outpatient setting) who have a high risk of VTE recurrence, we stratify according to the risk of thromboembolism progression or recurrence.

•Higher risk for VTE progression or recurrence – A higher risk for VTE progression or recurrence is suggested by an acute presentation (new VTE within the prior 30 days) and a proximal symptomatic lower extremity DVT or segmental or more central PE. For these individuals, we make an effort to treat with full-dose anticoagulation with platelet transfusion support (typically, platelet transfusions to raise the platelet count to ≥50,000/microL), especially in the presence of another thromboembolic risk factor. If the platelet count is low due to immune thrombocytopenia or certain other conditions, there may be other means of increasing the platelet count. (See 'Platelet count support' above.)

The rationale is that these individuals are more likely to have an adverse outcome related to thrombosis rather than bleeding, provided the platelet count can be consistently raised above 50,000/microL. However, this decision tree does not take the place of clinical judgment for the individual patient. If the platelet count cannot be consistently raised above 50,000/microL, or if the individual places an especially high value on avoiding full-dose anticoagulation and/or avoiding platelet transfusions (eg, due to a previous adverse reaction or alloimmunization), it may be appropriate to use another approach such as reduced-dose anticoagulation or temporarily holding anticoagulation, perhaps with the use of an IVC filter. (See "Placement of vena cava filters and their complications".)

When this transfusion strategy is used in patients with initial platelet counts <25,000/microL, and the count improves to the range of 25,000 to 50,000/microL, we consider using reduced dose anticoagulation and continuing the transfusion strategy as planned, since no anticoagulation can usually be given at counts below approximately 25,000/microL. We generally try to administer some anticoagulation, even at reduced doses whenever possible.

•Lower to intermediate risk for VTE progression or recurrence – A lower to intermediate risk for VTE progression or recurrence is suggested by an isolated distal DVT, isolated subsegmental PE, central-line associated DVT, or a subacute presentation (ie, >30 days since the acute VTE). For these individuals, we further stratify according to the platelet count.

-If the platelet count is <25,000/microL, we temporarily hold anticoagulation, with the plan to re-evaluate once the platelet count increases.

-If the platelet count is between 25,000 and 50,000/microL, we reduce the anticoagulant dose by one-half (eg, give enoxaparin 0.5 mg/kg twice daily rather than 1 mg/kg twice daily; give dalteparin 100 units/kg once daily rather than 200 units/kg once daily [applies to the initial 30 days of dosing]) [24]. Other adjustment schedules have also been used, such as reduction of the dalteparin dose by 2500 units as done in the Hokusai VTE cancer trial (see 'Supporting evidence' below). Dose reductions for unfractionated heparin have not been described, and we favor use of low molecular weight heparin (LMWH), but if unfractionated heparin is used, it would be reasonable to titrate to the lower end of the target activated partial thromboplastin time (aPTT) range and to avoid a bolus "loading" dose in these individuals.

The rationale is that in these individuals, the lack of the strongest risk factors for VTE progression or recurrence results in a shift of the balance towards a slightly greater concern about bleeding, and thus we do not use full-dose anticoagulation. At the same time, the risk for VTE progression or recurrence is still increased, and thus we still favor anticoagulation at a lower dose if possible. Some individuals who place a higher value on avoiding complications related to VTE progression or recurrence may choose full-dose (therapeutic-dose) anticoagulation with platelet support, and those who place a higher value on avoiding bleeding may choose to hold anticoagulation temporarily.

●With increased bleeding risk factors – For individuals who have one or more additional bleeding risk factors (age >75 years, recent severe bleeding, hematopoietic stem cell transplantation, kidney or liver failure, or increased risk for falls in the outpatient setting) along with a platelet count <50,000/microL, we generally do not use full-dose (therapeutic-dose) anticoagulation. Options include reducing the anticoagulant dose by one-half as noted above, using prophylactic rather than therapeutic dosing or temporarily holding anticoagulation, perhaps with the use of an inferior vena cava (IVC) filter. (See "Placement of vena cava filters and their complications".)

Our approach is generally consistent with a 2018 guideline from the International Society on Thrombosis and Haemostasis (ISTH) [6]. Supporting evidence is presented below. (See 'Supporting evidence' below.)

Exceptions may apply to certain populations such as individuals undergoing allogeneic hematopoietic stem cell transplantation, those with central venous catheters, and certain specific malignancies (brain tumors, multiple myeloma, pancreatic cancer). (See 'Special populations' below.)

The choice of anticoagulant in individuals with cancer is evolving, but the greatest experience is with unfractionated heparin and/or LMWH, as discussed separately. (See "Anticoagulation therapy for venous thromboembolism (lower extremity venous thrombosis and pulmonary embolism) in adult patients with malignancy", section on 'First-line options'.)

Certain direct oral anticoagulants (apixaban, edoxaban and rivaroxaban) have been shown to be effective in individuals with cancer, but these agents are generally associated with an increased risk of gastrointestinal hemorrhage, especially in patients with unresected gastrointestinal tumors. This is highlighted by a post-hoc analysis of patients in the HOKUSAI VTE cancer study (a randomized phase III trial comparing dalteparin with edoxaban in patients with cancer-associated VTE who had platelet counts below 100,000/microL), which demonstrated an increase in major bleeding with edoxaban in the subgroup of patients with gastrointestinal tumors (16.8 percent versus 0; p<0.01) [3].

Special populations — Certain patient populations deserve special consideration due to especially high risk of bleeding or thrombosis or the possibility of other options for treatment:

●Hematopoietic stem cell transplantation – In individuals undergoing hematopoietic stem cell transplantation, especially allogeneic, the risk of serious bleeding appears to be higher than the risk of serious complications from VTE. Thus, for individuals undergoing allogeneic transplantation who develop a VTE while thrombocytopenic, we favor dose-modified anticoagulation rather than therapeutic anticoagulation with aggressive platelet transfusion support.

This approach is supported by a retrospective cohort study of 250 patients undergoing allogeneic transplantation and 221 patients undergoing autologous transplantation, which demonstrated that continuing anticoagulant therapy (versus temporarily holding it) appeared to be associated with an increased risk of bleeding but was not associated with lower VTE risk within the first 30 days after transplantation [25]. In a subgroup of 250 allogeneic hematopoietic stem cell transplant patients at both intermediate and high risk of VTE, anticoagulation resumption upon platelet engraftment was associated with a lower risk of recurrent VTE up to 100 days (odds ratio [OR] 0.48, 95% CI 0.20-1.14). This emphasizes the importance of reinstating anticoagulation after platelet recovery (>50,000/microL for full-dose anticoagulation and >25,000/microL for reduced dose), if the initial indication still exists.

The risks of bleeding and VTE complications in individuals undergoing hematopoietic stem cell transplantation were illustrated in a 2008 series of 1514 individuals undergoing inpatient hematopoietic stem cell transplantation that identified 75 cases of symptomatic VTE in 70 patients (4.6 percent) within the first 180 days [26]. Most were catheter-associated (55 of 75; 73 percent). Approximately one-third occurred at a platelet count of <50,000/microL; the median platelet count at the time VTE developed was 75,000/microL. In a multivariate analysis, the features most predictive for VTE were prior VTE (odds ratio [OR] 2.9, 95% CI 1.3-6.6) and graft-versus-host disease (OR 2.4, 95% CI 1.4-4.0). There were no fatal VTE events. Clinically significant bleeding occurred in 230 individuals (15 percent) and fatal bleeding in 55 (4 percent).

●Catheter-associated VTE – Management of catheter-associated VTE may involve anticoagulation, thrombolysis within the catheter, or catheter removal. (See "Catheter-related upper extremity venous thrombosis in adults".)

When anticoagulation is withheld because of bleeding risk in patients with symptomatic catheter-associated VTE, catheter removal may be an option, after assessing the necessity of central venous access and alternatives [27,28]. Nonetheless, there is still a risk of recurrent thrombosis despite catheter removal. Therefore, this strategy should only be considered when anticoagulation cannot be given, rather than as an alternative to anticoagulation [27]. Reduced-dose anticoagulation has been used in individuals at risk for catheter-associated VTE [24]; this is a reasonable approach to prevent progression, as discussed above. (See 'Platelet count <50,000/microL' above.)

●Brain tumors – Management of VTE in an individual with a brain tumor (primary or metastatic) is discussed separately. (See "Treatment and prevention of venous thromboembolism in patients with brain tumors".)

●Multiple myeloma – Management of VTE in an individual with multiple myeloma is discussed separately. (See "Multiple myeloma: Prevention of venous thromboembolism".)

●Pancreatic cancer – Management of VTE in an individual with pancreatic cancer is discussed separately. (See "Supportive care for locally advanced or metastatic exocrine pancreatic cancer".)

Supporting evidence — Evidence regarding outcomes of cancer-associated VTE in individuals with concomitant thrombocytopenia comes from observational (often retrospective) studies. There are no randomized trials comparing different approaches to reducing the risks of VTE or VTE progression in people with cancer and thrombocytopenia.

The following studies illustrate some of the larger studies that have reported on outcomes of anticoagulation with different management strategies in patients who have cancer and thrombocytopenia. These generally describe successful anticoagulation for VTE, with dose adjustments for more severe thrombocytopenia; bleeding complications varied across the studies, while recurrent thrombosis appeared to be low in patients with reduced dose anticoagulation in some studies but not in others.

●A 2022 prospective observational study of 105 patients with hematologic malignancy, VTE within 28 days, and platelet count <50,000/microL demonstrated heterogeneity of management approaches and high rates of bleeding at 28 days (7 percent major bleeding and 25 percent clinically relevant nonmajor bleeding) [2]. There was a high rate of thrombus progression at 28 days (11 to 13 percent), which was similar in patients receiving full- and reduced-dose LMWH.

●A 2021 prospective observational study involving 121 patients with solid tumors and hematologic malignancies, acute VTE, and thrombocytopenia (platelet count <100,000/microL) found better 60-day outcomes with modified-dose anticoagulation rather than full therapeutic dosing [1]. The median platelet count was 65,000/microL in the full-dose group and 37,000/microL in patients receiving modified doses. With modified dosing (mostly used in individuals with hematologic malignancies), the rate of VTE recurrence was 0 and major bleeding was 6.6 percent (95% CI 2.4-15.7). With full therapeutic dosing, the rate of VTE recurrence was 5.6 percent (95% CI, 0.2-11) and major bleeding was 12.8 percent (95% CI 4.9-20.8).

●A 2020 study analyzed 166 patients in the RIETE registry with active cancer, acute VTE and platelet count <50,000/microL and found that the rates of major bleeding at 30 days were similar in those with reduced-dose and therapeutic-dose LMWH (3.4 percent and 2.9 percent, P = 0.86) [29]. In contrast, VTE recurrence rate was not significantly higher in patients on reduced LMWH doses relative to therapeutic LMWH doses (10.3 versus 1.4 percent, P = 0.08).

●A 2017 quality initiative project identified 99 individuals with cancer who were receiving therapeutic-dose enoxaparin for VTE and had at least one episode of thrombocytopenia lasting seven or more days (median duration, 12 days) over a three-year period [30]. This represented approximately 0.6 percent of individuals with VTE during the study period. Most had a hematologic malignancy (59 percent) and most had been treated with chemotherapy within the prior month (79 percent). LMWH was used at full therapeutic dosing for those with platelet counts >50,000/microL, reduced to half-dose for those with platelet counts between 25,000 and 50,000/microL, and temporarily held for a platelet count <25,000/microL. Over the course of the project, 95 percent of the patients had some form of LMWH dose adjustment or temporary discontinuation. There were no instances of recurrent VTE or major bleeding in patients treated according to this approach. One individual receiving full-dose enoxaparin with a platelet count of 28,000/microL had a traumatic retroperitoneal hemorrhage.

These somewhat conflicting findings may be explained in part by different patient populations but also emphasize the need for prospective randomized trials.

OTHER SELECTED CLINICAL SCENARIOS — There are other clinical settings in which thrombocytopenia coexists with an increased risk of thrombosis for which anticoagulation should be administered. In some cases, two common conditions, one requiring anticoagulation and one causing thrombocytopenia, may be present (eg, atrial fibrillation plus immune thrombocytopenia). Less common causes include antiphospholipid syndrome, especially with immune thrombocytopenia; PNH; and heparin-induced thrombocytopenia.

Liver disease — Liver disease has multiple effects on coagulation proteins, resulting in a "rebalanced" hemostasis. It used to be stated that individuals with liver disease were auto-anticoagulated due to reduced levels of certain procoagulant factors made in the liver. However, better documentation of the multiple effects of liver disease on hemostasis has made it clear that this is not the case; thrombosis and bleeding risk are both increased in the same patient (rates of deep vein thrombosis [DVT] and portal vein thrombosis are increased, and cirrhosis is associated with thrombocytopenia and coagulation abnormalities). Management of venous thromboembolism (VTE) in individuals with liver disease is discussed separately. (See "Hemostatic abnormalities in patients with liver disease", section on 'Portal vein thrombosis (PVT)' and "Hemostatic abnormalities in patients with liver disease", section on 'Venous thromboembolism (VTE)'.)

Atrial fibrillation — Atrial fibrillation is common and often occurs in the setting of another disorder that affects platelet counts such as myelodysplastic syndrome, cancer or immune thrombocytopenia. While the bleeding risk associated with anticoagulation increases in thrombocytopenic patients, the disorders causing the thrombocytopenia are often also associated with an increased thrombotic risk.

●In a population-based cohort study of 1411 patients with cancer, atrial fibrillation, and a CHA₂DS₂-VASc score of 0 to 2, not receiving anticoagulation, the risk of stroke was increased relative to 4233 individuals with atrial fibrillation who did not have cancer (hazard ratio [HR] 2.70, 95% CI 1.65-4.41) [10]. The patients with cancer also had an increased bleeding risk (HR 2.79, 95% CI 1.39-5.58), emphasizing the delicate balance between bleeding and thrombosis in patients with cancer and atrial fibrillation.

●A retrospective cohort study of 274 patients with atrial fibrillation receiving anticoagulation demonstrated a higher incidence of major and clinically relevant nonmajor bleeding at 12 months in those with platelet counts <100,000/microL versus normal platelet counts (>150,000/microL), but there was no increase in risk of arterial thrombosis [4]. Bleeding was the dominant outcome at 12 months, with a 13.3 percent incidence of major bleeding and 24.5 percent clinically relevant nonmajor bleeding, while only 3.6 percent had arterial thrombosis. Only 22 percent of the thrombocytopenic patients in this study had cancer. A small retrospective study of patients with atrial fibrillation, leukemia, and platelet counts <50,000/microL demonstrated a similar risk-benefit ratio [31].

There is no strong evidence that reduced-dose anticoagulation prevents stroke in atrial fibrillation patients with thrombocytopenia. Therefore, the decision should generally be between continuing anticoagulation or holding anticoagulation temporarily or permanently.

The anticipated duration of thrombocytopenia is an important consideration, since a thrombocytopenia duration of days to weeks may carry a low absolute risk of arterial thrombosis and support temporary holding of anticoagulation. Due to the high bleeding risk, the thrombotic risk should be high enough to justify the increased bleeding risk with anticoagulation, such as with recent stroke or high CHA₂DS₂-VASc score (eg, ≥4) without additional bleeding risk factors [5].

When anticoagulation is discontinued permanently due to thrombocytopenia in patients with high thrombotic risk, left atrial appendage occlusion should be considered on a case-by-case basis, in patients without or with cancer [5,32].

Disorders that simultaneously cause thrombosis and thrombocytopenia — For most of these disorders, there are treatments that can be directed at the underlying cause of the thrombocytopenia (eg, immunosuppressive therapy for systemic lupus erythematosus, anti-complement therapy in PNH).

If these individuals develop thrombosis, therapy for the underlying disorder may lessen bleeding and thrombotic risks. In the interim (while waiting for disease therapy to take effect), it may be necessary to transfuse platelets to facilitate anticoagulation. (See 'Platelet count support' above.)

Some of these disorders are discussed in the following sections; management of disseminated intravascular coagulation is presented separately. (See "Evaluation and management of disseminated intravascular coagulation (DIC) in adults".)

Antiphospholipid syndrome — The antiphospholipid syndrome is characterized by venous or arterial thrombosis and/or adverse pregnancy outcomes in the presence of persistent autoantibodies directed against phospholipid-binding proteins. It can occur as an isolated syndrome or in the setting of systemic lupus erythematosus or another autoimmune disorder. The antiphospholipid antibodies may artificially prolong the activated partial thromboplastin time or, less commonly, the prothrombin time, but in vivo they promote thrombosis. (See "Clinical manifestations of antiphospholipid syndrome", section on 'Thrombotic events'.)

Thrombocytopenia is common in individuals with antiphospholipid syndrome, which may be isolated or associated with systemic lupus erythematosus (eg, as a form of secondary immune thrombocytopenia), with a prevalence of approximately 20 to 40 percent in some series. Typically, the thrombocytopenia is mild (platelet counts 100,000 to 149,000/microL), although more severe thrombocytopenia is seen in some individuals, especially those with systemic lupus erythematosus and/or immune thrombocytopenia. (See "Clinical manifestations of antiphospholipid syndrome", section on 'Hematologic abnormalities' and "Hematologic manifestations of systemic lupus erythematosus", section on 'Thrombocytopenia'.)

For most patients with thrombotic antiphospholipid syndrome who have platelet counts >50,000/microL and are otherwise considered lower risk for hemorrhage, the preferred anticoagulant is warfarin. (See "Management of antiphospholipid syndrome", section on 'Approach to anticoagulation'.)

Thrombocytopenia is not protective against cardiovascular or VTE events related to antiphospholipid syndrome. Our approach in individuals with antiphospholipid syndrome and concurrent immune thrombocytopenia is to administer immune thrombocytopenia-directed therapies to rapidly raise the platelet count above 50,000/microL to facilitate therapeutic anticoagulation. This may include the use of intravenous immune globulin, glucocorticoids (eg, high-dose dexamethasone), and/or a thrombopoietin receptor agonist (TPO-RA). While these therapies may carry an increased risk of thrombosis, they would be administered to enable anticoagulation, which should be initiated at an appropriate dose according to platelet counts. (See "Initial treatment of immune thrombocytopenia (ITP) in adults".)

At lower platelet counts, we follow similar approach to that used in cancer-associated thrombosis (algorithm 1), with the one caveat being that platelet transfusion is generally less effective in raising the platelet count and is generally not used unless there is acute bleeding.

PNH — Paroxysmal nocturnal hemoglobinuria (PNH) is an acquired abnormality characterized by expansion of an abnormal clone of hematopoietic cells lacking a glycosylphosphatidylinositol anchor, which leads to hemolysis and thrombosis. The mechanism of thrombosis is incompletely understood and probably involves release of free hemoglobin, nitric oxide scavenging, vasoconstriction, and endothelial cell activation. Many individuals with PNH also have aplastic anemia, which can result in severe thrombocytopenia. (See "Pathogenesis of paroxysmal nocturnal hemoglobinuria".)

PNH with severe hemolysis can be treated with complement inhibitors such as eculizumab, ravulizumab, or pegcetacoplan; severe aplastic anemia may require hematopoietic cell transplantation. As with other disorders, thrombocytopenia is not protective against thrombosis, and individuals with thrombosis are treated with anticoagulation. (See "Treatment and prognosis of paroxysmal nocturnal hemoglobinuria".)

Heparin-induced thrombocytopenia — Heparin-induced thrombocytopenia is an adverse drug reaction to heparin in which antibodies to platelet factor 4 (PF4) in a complex with heparin lead to thrombocytopenia and a dramatically increased risk of venous and arterial thrombosis. Management of heparin-induced thrombocytopenia involves scrupulous avoidance of all heparin-containing products and anticoagulation with a non-heparin anticoagulant. Typically, thrombocytopenia is moderate (eg, mean nadir platelet count approximately 60,000/microL) and resolves rapidly when heparin is discontinued, although some individuals may have more prolonged thrombocytopenia. (See "Clinical presentation and diagnosis of heparin-induced thrombocytopenia", section on 'Thrombocytopenia'.)

Despite thrombocytopenia, management of heparin-induced thrombocytopenia (even in the absence of thrombosis) requires full-dose anticoagulation with a non-heparin anticoagulant and the expectation that discontinuation of heparin exposure will correct the thrombocytopenia, as discussed in detail separately. (See "Management of heparin-induced thrombocytopenia" and "Management of heparin-induced thrombocytopenia (HIT) during cardiac or vascular surgery".)

SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Anticoagulation".)

SUMMARY AND RECOMMENDATIONS

●Deciding to use anticoagulation – Decisions about anticoagulation in patients with thrombocytopenia must balance the risks of thrombosis and thrombosis progression with the risks of bleeding due to thrombocytopenia as well as other bleeding risk factors. Risk factors for bleeding (other than thrombocytopenia) and thrombosis are listed above, along with other factors that may contribute to decision-making. (See 'General principles of treatment' above.)

●Malignancy – Cancer-associated venous thromboembolism (VTE) is one of the most common scenarios in which VTE and thrombocytopenia coexist. (See 'Cancer-associated VTE' above.)

•Prophylaxis – For individuals with platelet counts of 50,000/microL or above, anticoagulation for VTE prophylaxis can be done according to established protocols, which are discussed in separate topic reviews. (See "Prevention of venous thromboembolic disease in acutely ill hospitalized medical adults" and "Risk and prevention of venous thromboembolism in adults with cancer" and "Prevention of venous thromboembolic disease in adult nonorthopedic surgical patients".)

For those with counts between 25,000 and 50,000/microL, VTE prophylaxis is individualized, and for counts below 25,000/microL, prophylactic anticoagulation generally is not used. (See 'Primary VTE prophylaxis' above.)

•Treatment – For individuals with platelet counts of 50,000/microL or above, anticoagulation for VTE treatment can be administered according to established protocols. (See 'Platelet count 50,000/microL or above' above.)

For most individuals with cancer-associated VTE and platelet counts between 25,000 and 50,000/microL, we suggest anticoagulation (Grade 2C). The rationale is that morbidity from thrombosis is likely to be greater than morbidity from bleeding.

Options for dosing include the following:

-Higher risk of VTE progression/recurrence – Full (therapeutic)-dose anticoagulation with platelet support (eg, platelet transfusions to raise the platelet count above 50,000/microL). For most individuals with a high risk of VTE progression or recurrence (ie, acute symptomatic proximal lower extremity deep vein thrombosis or pulmonary embolism that is segmental or more proximal) who do not have additional bleeding risk factors other than thrombocytopenia, we suggest this approach (Grade 2C). In those patients whereby transfusion support is not practical or feasible, or when platelet counts remain in the 25,000 to 50,000/microL range, we use half-dose anticoagulation. In such patients with additional bleeding risk factors who cannot receive anticoagulation, placement of an inferior vena cava (IVC) filter may be considered on a case-by-case basis.

-Lower to intermediate risk of VTE progression/recurrence or high risk of bleeding – Half-dose anticoagulation (eg, enoxaparin 0.5 mg/kg twice daily rather than 1 mg/kg twice daily) is the preferred treatment approach for platelet counts between 25,000 to 50,000/microL; holding anticoagulation is appropriate for platelet counts <25,000/microL. Other options include prophylactic-dose anticoagulation or temporarily holding anticoagulation, possibly with removal of the central venous catheter in case of catheter-related DVT. For most individuals with a low risk for VTE progression and/or those with one or more additional bleeding risk factors besides thrombocytopenia, we suggest one of these approaches (Grade 2C).

Supporting evidence comes from observational studies. (See 'Supporting evidence' above.)

Individuals with VTE may reasonably make different choices depending on their specific circumstances, bleeding risk factors, and VTE recurrence risk, as illustrated in the algorithm (algorithm 1). (See 'Platelet count <50,000/microL' above.)

Platelet transfusions are accepted therapy to raise the platelet count if needed in order to provide full-dose anticoagulation. In contrast, we do not routinely use thrombopoietin receptor agonists (TPO-RAs) as a means of facilitating anticoagulation outside of accepted indications such as immune thrombocytopenia. (See 'Platelet count support' above.)

Additional considerations may apply to individuals undergoing hematopoietic cell transplant, those with a central venous catheter, and certain malignancies such as multiple myeloma, pancreatic cancer, or brain tumors. (See 'Special populations' above.)

●Other conditions – Other disorders that are commonly associated with thrombocytopenia plus an indication for anticoagulation include liver disease, atrial fibrillation, antiphospholipid syndrome, paroxysmal nocturnal hemoglobinuria, and heparin-induced thrombocytopenia. Our approaches to each of these conditions is discussed above and/or in the linked topic reviews. (See 'Other selected clinical scenarios' above.)