Clinical presentation and diagnosis of heparin-induced thrombocytopenia

INTRODUCTION — Heparin-induced thrombocytopenia (HIT) is a life-threatening complication of exposure to heparin (eg, unfractionated heparin, low molecular weight [LMW] heparin) that occurs in a small percentage of patients exposed, regardless of the dose, schedule, or route of administration.

HIT results from an autoantibody directed against endogenous platelet factor 4 (PF4) in complex with heparin. This antibody activates platelets and can cause catastrophic arterial and venous thrombosis. Untreated HIT has a mortality rate as high as 20 percent; although with improved recognition and early intervention, mortality rates have been reported as below 2 percent.

The clinical presentation and diagnosis of HIT will be discussed here. The management of HIT, HIT in patients undergoing cardiopulmonary bypass, the related syndrome of coronavirus disease 2019 vaccine-induced thrombosis and thrombocytopenia (VITT), and details of administration of direct oral anticoagulants (DOACs) and warfarin are discussed separately.

●Management of HIT – (See "Management of heparin-induced thrombocytopenia".)

●HIT during cardiopulmonary bypass – (See "Management of heparin-induced thrombocytopenia (HIT) during cardiac or vascular surgery".)

●VITT – (See "COVID-19: Vaccine-induced immune thrombotic thrombocytopenia (VITT)".)

●DOACs – (See "Direct oral anticoagulants (DOACs) and parenteral direct-acting anticoagulants: Dosing and adverse effects".)

●Warfarin – (See "Warfarin and other VKAs: Dosing and adverse effects".)

TERMINOLOGY AND HIT VARIANTS

●HIT type I versus type II – There are two forms of HIT, only one of which is clinically significant (table 1). Distinction between these two forms of HIT is made based on clinical parameters such as the timing and degree of platelet count drop, and laboratory testing if needed.

•HIT type I (HIT I) is a mild, transient drop in platelet count that typically occurs within the first two days of heparin exposure. The platelet count typically returns to normal with continued heparin administration. The mechanism appears to be a direct effect of heparin on platelets, causing non-immune platelet aggregation. The typical platelet count nadir is approximately 100,000/microL. This form of HIT is not considered clinically significant, is not associated with thrombosis, and patients can be managed expectantly without discontinuation of heparin.

•HIT type II (HIT II) is a clinically significant syndrome due to antibodies to platelet factor 4 (PF4) complexed to heparin, referred to as "HIT antibodies" or "PF4/heparin antibodies" [1]. These antibodies can cause thrombosis along with thrombocytopenia; hence, this syndrome has also been called heparin-induced thrombocytopenia and thrombosis (HITT). The risk of thrombosis, including life-threatening limb gangrene, persists until both heparin is eliminated and a non-heparin anticoagulant is initiated.

●Phases of HIT/subclinical HIT – Phases of HIT have been defined that include the acute condition as well as various phases of recovery (table 2) [2,3]. Subclinical (also subacute) HIT refers to the state where a patient has recovered from an episode of HIT (ie, their platelet count has returned to normal or the patient's baseline) but has persistent HIT antibodies. Remote HIT refers to platelet count recovery and negative heparin-PF4 antibodies. Individuals with subacute or remote HIT are at high risk of HIT recurrence if re-exposed to heparin. Subclinical HIT should be distinguished from the presence of naturally occurring or clinically silent anti-PF4 antibodies (eg, with no associated thrombocytopenia, thrombosis, or other clinical manifestations).

●HIT variants – Rarely, clinical HIT may be present due to HIT antibodies that activate platelets in the absence of heparin [4]. These syndromes are sometimes referred to as autoimmune HIT (table 3). Examples include the following:

•Delayed-onset HIT – Delayed-onset HIT is defined by thrombocytopenia and/or thrombosis occurring five or more days after heparin has been withdrawn. This presentation may be related to high-titer antibodies to PF4/heparin [5]. (See 'Delayed-onset HIT following withdrawal of heparin' below.)

•Refractory (persistent) HIT – Refractory HIT is HIT with persistent thrombocytopenia and/or thrombosis that lasts for weeks after stopping heparin [4]. The presence of HIT antibodies alone does not define this condition.

•Spontaneous HIT – Rarely, HIT has been described in the absence of recent heparin exposure; this condition has been termed "spontaneous HIT." Most affected patients have a preceding infection or have undergone a major surgical procedure. The following diagnostic criteria have been proposed for this disorder [6]:

-Otherwise unexplained thrombocytopenia and/or thrombosis without recent heparin exposure

-Demonstration of anti-PF4 antibodies of the immunoglobulin G (IgG) subclass that cause strong in vitro platelet activation in the absence of heparin

-Occurring in the setting of exposure to minimal amount of heparin, as in "heparin flushes"

Spontaneous HIT likely represents a form of autoimmune reaction to endogenous heparins [6].

•VITT – Vaccine-induced immune thrombotic thrombocytopenia (VITT, also called thrombosis with thrombocytopenia syndrome [TTS] or vaccine-induced prothrombotic immune thrombocytopenia [VIPIT]), is a rare complication of adenoviral vectored coronavirus disease 2019 (COVID-19) vaccines including the ChAdOx1 CoV-19 vaccine (AstraZeneca, University of Oxford, and Serum Institute of India) and the Ad26.COV2.S vaccine (Janssen; Johnson & Johnson). Evaluation and management of VITT are discussed separately (See "COVID-19: Vaccine-induced immune thrombotic thrombocytopenia (VITT)".)

●Heparin-induced antibodies (HIA) – Some patients produce antibodies that react in laboratory assays for HIT but do not cause thrombocytopenia or thrombosis. These antibodies can occur in patients exposed to heparin (eg, during cardiopulmonary bypass) or with other conditions in the absence of heparin exposure (eg, lupus) and are sometimes referred to as heparin-induced antibodies (HIA). (See 'Formation of HIT antibodies' below.)

PATHOPHYSIOLOGY

Formation of HIT antibodies — HIT has an unusual immunobiology compared with other drug-induced antibodies, which includes the following features [7]:

●IgG antibodies are formed rapidly (within days). There is no primary IgM response for clinically significant platelet factor 4 (PF4)/heparin antibodies.

●IgG antibodies disappear over time; however, this does not imply that heparin can be used upon disappearance of the antibodies. (See "Management of heparin-induced thrombocytopenia", section on 'Lifelong heparin avoidance (list of sources)'.)

●Immune memory does not always occur (ie, rechallenge does not always cause an anamnestic response), although individuals with recent heparin exposure may develop clinical findings of HIT in a shorter timeframe than those without recent heparin exposure. (See 'Timing' below.)

HIT antibodies are directed against PF4 complexed with heparin (figure 1) [1,8-11]. It is thought that the binding of heparin (or other polyanions) induces a conformational change in the PF4 protein that creates a neoantigen, and HIT antibodies bind to this PF4 neoantigen [12-14]. As a result of this heparin dependency, these antibodies usually only cause clinical symptoms when heparin is present. Initial formation of HIT antibodies typically requires exposure to heparin for four or more days, with rare exceptions listed above.

Anti-heparin-PF4 antibodies can be of the IgG, IgM, and IgA subclass. IgG is thought to be the only pathogenic antibody because the platelet surface Fc receptor only recognizes IgG.

Naturally occurring antibodies that react with PF4 bound to heparin are present in 3 to 8 percent of the general population without HIT [15,16]. The prevalence of high-titer antibodies is significantly lower (0.3 percent with an optical density >1.0 in one study) [16]. It has been suggested that these antibodies are induced when PF4 binds to negatively charged polysaccharides on the surface of bacteria, creating an antigen in PF4 that mimics PF4 bound to heparin [15]. This pre-exposure to the PF4 antigen may be responsible for the rapid appearance of HIT antibodies following treatment of some patients with heparin.

The HIT immune response is similar to the T-cell-independent B-cell activation that is seen with immune reactions against viruses with repetitive epitopes [7]. Preliminary evidence suggests that this may be the explanation for the syndrome seen with adenoviral vector COVID-19 vaccines. (See 'Terminology and HIT variants' above.)

Platelet factor 4 (PF4) — The PF4 protein is stored in platelet alpha granules. It is small cytokine belonging to the CXC chemokine family. Platelet activation causes release of PF4, which forms a tetramer that binds to and neutralizes heparin and related endogenous proteoglycan molecules (eg, heparan sulfate, chondroitin sulfate) on endothelial surfaces [17]. Heparin-PF4 complexes also form on the platelet surface.

HIT antibodies may bind to heparin-PF4 complexes on the platelet surface (via heparin binding to cell surface proteins) or to PF4 attached to platelet surface glycosaminoglycans (heparin-like molecules) [8,18]. In an in vitro assay, HIT antibodies were shown to also bind to PF4 complexed with high molecular weight multimer strings of von Willebrand factor (VWF) released from injured endothelium under flow conditions, which also promoted platelet binding and activation [19].

Once HIT antibodies bind to PF4 on the platelet surface, their Fc region is captured by Fc receptors on the surface of the same or adjacent platelets (Fc gamma Receptor IIA [FcγRIIA]), as well as glycoprotein (GP) Ib/IX, creating a positive feedback loop of further platelet activation [20]. This in turn leads to release of more PF4, which creates more antigenic substrate for HIT antibodies [21].

Mechanism of thrombocytopenia — The mechanisms of thrombocytopenia in HIT include removal of IgG-coated platelets by macrophages of the reticuloendothelial system (eg, spleen, liver, bone marrow), via binding to the FcγRIIA, similar to other types of drug-induced immune thrombocytopenia; consumption of platelets at sites of thrombosis; and platelet destruction due to the development of a consumptive coagulopathy. A second cause of thrombocytopenia is consumption of platelets within thrombi.

Mechanism of thrombosis

Platelet activation and endothelial cell injury — Unlike most other forms of drug-induced immune thrombocytopenia, HIT is also associated with arterial and venous thrombosis. The mechanism is likely multifactorial, but the primary mechanism is thought to be due to activation of platelets and injury to endothelial cells via activation of the cell surface FcγRIIA on platelets and endothelial cells, respectively. This results in the following downstream effects (figure 1):

●Release of procoagulant substances, including microparticles, from activated platelets [22]

●Endothelial cell activation and/or injury induced by binding of HIT antibodies to heparan sulfate on endothelial cell surfaces, leading to increased tissue factor expression and thrombin generation [23,24]

●Endothelial cell release of adhesion molecules (eg, interleukin-6, von Willebrand factor) [25]

●Activation of monocytes by HIT antibodies [26-28]

●Alteration of other aspects of coagulation by HIT antibodies (eg, reduced generation of activated protein C) [29]

A genetic polymorphism in the platelet-associated immunoglobulin Fc receptor has been proposed to enhance platelet activation by increasing the binding of HIT antibodies. In a study of 89 patients with HIT, thrombosis was more frequent in individuals with an arginine rather than a histidine at amino acid position 131 on both alleles of the receptor (ie, FcγRIIA 131RR rather than FcγRIIA 131HH) [30]. The odds ratio (OR) for thromboembolic events with the 131RR allele was 5.9 (95% CI 1.7-20).

Neutrophil activation and NETosis — Neutrophil extracellular traps (NETs) are structures created when neutrophils extrude their DNA as uncoiled strands of chromatin that incorporate cellular contents such as histones, myeloperoxidase, and elastase. NETs play a central role in host defense against infection. (See "An overview of the innate immune system", section on 'Neutrophils'.)

Studies suggest that NETosis also plays a key role in thrombosis in individuals with HIT [31,32].

HIT immune complexes can activate neutrophils by two pathways:

●Directly, via FcγRIIA

●Indirectly, by activation of platelets via FcγRIIA followed by interaction and activation of neutrophils via P-selectin and its ligand P-selectin glycoprotein ligand-1 (PSGL-1)

In a study involving 21 individuals with HIT, plasma markers of NETosis including cell-free DNA, myeloperoxidase, "low density granulocytes," and citrullinated histone H3 were increased substantially compared with control plasma [31]. Plasma from HIT patients and purified antibodies to heparin-PF4 were able to induce release of DNA from purified neutrophils. In a mouse model of HIT, depletion of neutrophils or blockage of nuclear decondensation in neutrophils by a NETosis inhibitor prevented thrombosis, while thrombocytopenia was unaffected, supporting the important role of NETosis in HIT thrombosis. The data also suggest that thrombosis and thrombocytopenia in HIT are separable events.

Clinical variability — The clinical presentation of HIT (eg, risk of thrombosis, duration of antibody persistence) varies in different patient populations. While many of the mechanisms are unknown, increases in systemic inflammation and differences in platelet surface glycoproteins may be involved [33-38]. The greater incidence of HIT in women is unexplained. (See 'Incidence and risk factors' below.)

HIT can occur with any heparin dose, schedule, and administration route. However, the molar ratio of PF4 to heparin influences the concentration and size of PF4-heparin complexes on the platelet surface, which in turn influences antigenicity and clinical sequelae. A molar ratio of PF4 tetramers to heparin molecules of 1:1 causes formation of more ultralarge complexes and is more immunogenic [10].

Ultralarge complexes form most efficiently with unfractionated heparin; their formation is 10-fold less efficient with low molecular weight (LMW) heparin and negligible with fondaparinux. These biochemical findings correlate with the greater incidence of HIT in patients receiving unfractionated compared with LMW heparin [39]. Patients with greater amounts of PF4 in the circulation or on platelet surfaces also have a higher likelihood and/or severity of HIT [12,33].

The optimal 1:1 molar ratio for forming ultralarge PF4-heparin complexes may also explain why those exposed to very high doses of heparin (eg, during cardiopulmonary bypass) are less likely to develop HIT than those exposed to standard doses. This stoichiometric dependence also underscores the basis for using high doses of heparin to demonstrate reduced binding of true HIT antibodies in immunoassays for HIT (See 'Cardiac surgery/cardiopulmonary bypass' below and 'Immunoassay (eg, ELISA)' below and "Management of heparin-induced thrombocytopenia (HIT) during cardiac or vascular surgery".)

The titer of the HIT antibody may also influence the timing of HIT. High-titer antibodies that react with PF4 bound to non-heparin glycosaminoglycans (eg, chondroitin sulfate) have been proposed to explain the occurrence of HIT in patients who have never been exposed to heparin or those who develop delayed-onset HIT following heparin withdrawal [12,40]. (See 'Delayed-onset HIT following withdrawal of heparin' below.)

INCIDENCE AND RISK FACTORS — HIT has been reported in up to 5 percent of patients exposed to heparin for more than four days [41-49]. Factors that increase the frequency of HIT include the following:

●Surgery – An analysis of seven trials comparing unfractionated versus low molecular weight (LMW) heparin found a greater incidence of HIT in surgical versus medical patients (odds ratio [OR] 3.25; 95% CI 1.98-5.35) [45].

Patients undergoing cardiac surgery or cardiopulmonary bypass have a very high incidence of anti-PF4/heparin antibody formation (eg, 20 to 50 percent). However, the risk of true HIT in these patients is similar to or lower than other surgical patients, as discussed in more detail separately. (See "Management of heparin-induced thrombocytopenia (HIT) during cardiac or vascular surgery", section on 'Overview and scope of the problem'.)

●Unfractionated versus LMW heparin – Patients can develop HIT regardless of whether their prior heparin exposure was to unfractionated or LMW heparin, but HIT occurs more commonly after exposure to unfractionated heparin than exposure to LMW heparin in surgical patients. A meta-analysis of 15 studies that evaluated the risk of HIT with prophylactic unfractionated versus LMW heparin (7287 patients, most undergoing orthopedic surgery) found the following absolute risks of developing HIT [41]:

•Unfractionated heparin – 2.6 percent (95% CI 1.5-3.8 percent)

•LMW heparin – 0.2 percent (95% CI 0.1-0.4 percent)

It is unclear if medical patients also have a lower risk of HIT with LMW than unfractionated heparin, as some studies have shown this effect and others have not [45-47]. A meta-analysis of 13 studies comparing unfractionated versus LMW heparin in 5275 medical patients with venous thromboembolism (VTE) found no difference in the risk of HIT according to the type of heparin used [44].

●Heparin dose – Therapeutic doses of heparin may result in a greater incidence of HIT than prophylactic doses, but there are no data that clearly define the relationship between heparin dose and clinical findings. In a hospital database review of 24,068 patients exposed to unfractionated heparin, HIT occurred in 0.76 percent receiving therapeutic intravenous heparin and <0.1 percent receiving prophylactic subcutaneous heparin [49].

However, there is no dose of heparin that is too low to cause HIT. Patients have developed HIT after exposure to as little as 250 units from a heparin flush or after the use of heparin-coated catheters [50-52]. In some cases, these may represent a form of autoimmune HIT, as strong platelet activation induced by patient serum has been demonstrated in the absence of heparin [53].

●Female sex – Analysis of seven trials comparing unfractionated versus LMW heparin found approximately twice the risk of HIT in female patients compared with males (OR 2.37; 95% CI 1.37-4.09) [45]. A higher incidence of HIT in women than in men was also found in analyses of a national database (807 patients) and a randomized trial of unfractionated versus LMW heparin (665 patients) analyzed by the same authors [45]. Of interest, this female predominance of HIT was restricted to those receiving unfractionated heparin (ie, the incidence of HIT was similar in females and males receiving LMW heparin). The highest risk for HIT was seen in female surgical patients receiving unfractionated heparin (incidences from 2.7 to 7.6 percent; OR 17; 95% CI 4.2-72) [45].

A similar predilection for females has been observed in vaccine-induced immune thrombotic thrombocytopenia (VITT)/vaccine-induced prothrombotic immune thrombocytopenia (VIPIT) associated with adenoviral coronavirus disease 2019 vaccines.

●Age – Older age may be a risk factor for HIT, but good data illustrating a relationship between age and HIT are lacking. In a study using the database of the National Hospital Discharge Survey, HIT was found to be rare among patients <40 years of age and in the postpartum period [46].

HIT is thought to be very rare in children and in patients receiving hemodialysis [54-57]. In one series that identified 34 children referred for HIT testing, most were exposed to unfractionated heparin during cardiopulmonary bypass or extracorporeal membrane oxygenation (ECMO) [57].

CLINICAL MANIFESTATIONS

Thrombocytopenia — Thrombocytopenia (platelet count <150,000/microL) is the most common manifestation of HIT, occurring in 85 to 90 percent of individuals [39,58]. The mean nadir platelet count is approximately 60,000/microL. Platelet counts below 20,000/microL are rare [39,59]. Approximately 5 percent of patients with HIT lack thrombocytopenia as defined by absolute platelet count but demonstrate a 50 percent reduction in platelet count [39].

Bleeding — Bleeding is uncommon but has been reported, sometimes in unusual sites [60-62]. In a series of 6332 patients hospitalized with HIT, bleeding was seen in approximately 6 percent [61]. The study did not report whether bleeding was due to thrombocytopenia from HIT or to anticoagulation for HIT treatment. Gastrointestinal bleeding was the most common type, seen in 2.7 percent of patients, and central nervous system bleeding was seen in approximately 1 percent.

Some forms of bleeding in HIT may be a dire manifestation of a thrombotic event, such as adrenal hemorrhage in patients with adrenal thrombosis or intracranial hemorrhage in patients with cerebral sinus thrombosis. In both cases, bleeding is thought to be due to venous engorgement and "back pressure" rather than due to a vascular lesion.

Timing

Typical presentation — HIT typically occurs 5 to 10 days after the initiation of heparin [63,64]. Heparin-dependent antibodies usually develop between five and eight days after heparin exposure [42,65]. The figure illustrates the typical timing, which provides the rationale for platelet count monitoring (figure 2).

Early onset of HIT (ie, thrombocytopenia within the first 24 hours of exposure) may be seen if the patient has been exposed to heparin in the previous one to three months and has circulating HIT antibodies [58,64]. In a review of 243 patients with HIT, early onset of thrombocytopenia (median: 10.5 hours after heparin initiation) occurred in 30 percent, all of whom had been exposed to heparin within the previous 100 days [63].

The resolution of thrombocytopenia following withdrawal of heparin and initiation of a non-heparin anticoagulant typically occurs within seven days. The possibility of continued exposure to heparin or an additional cause of thrombocytopenia should be investigated in a patient with HIT whose platelet count does not start to improve within three to four days of heparin withdrawal.

Delayed-onset HIT following withdrawal of heparin — Delayed-onset HIT is a well-described condition in which thrombocytopenia and/or thrombosis occur after heparin has been withdrawn; the incidence is unknown [5,66]. In one study, HIT occurred in 12 patients an average of nine days after heparin was withdrawn (range: 5 to 19 days) [5]. These patients had high-titer HIT antibodies that exhibited heparin-dependent as well as heparin-independent platelet activation, potentially explaining why complications occurred in the absence of heparin.

In a retrospective study from three hospitals, there were 14 cases of delayed-onset HIT occurring at a median time of 14 days after discharge from the hospital (range: 9 to 40 days) [66]. Most patients were exposed to heparin during the initial hospitalization and did not develop complications until after discharge from the hospital. On readmission, 11 of the 14 patients were inadvertently treated with heparin, resulting in a prompt decrease in platelet counts, often with overt clinical deterioration; three patients died. After HIT was recognized, the remainder were treated with non-heparin anticoagulants and ultimately recovered.

Thrombosis — Thrombosis occurs in up to 50 percent of individuals with HIT who are not treated with a non-heparin anticoagulant, with venous more common than arterial thrombi. Thrombosis is the presenting finding in up to 25 percent of patients; this observation has led to recommendations that patients with overt HIT have leg ultrasounds performed to screen for asymptomatic deep vein thrombosis [39]. If clinical signs or symptoms suggest venous or arterial thrombosis at other sites, these should be promptly investigated with appropriate testing.

The reported frequencies range from 20 to 50 percent for venous thrombosis and 3 to 10 percent for arterial thrombosis [39].

●A retrospective review of 127 patients with serologically confirmed HIT who were not treated with parenteral anticoagulation (ie, they were only treated with heparin cessation or warfarin) found that venous and arterial thrombosis occurred in 61 and 14 percent, respectively [67]. Of the patients initially recognized with isolated thrombocytopenia, the subsequent 30-day risk of thrombosis was 53 percent. Pulmonary embolism was the most common life-threatening event, occurring in 25 percent of patients.

●In a randomized trial that compared unfractionated heparin with LMW heparin for venous thromboembolism (VTE) prophylaxis in 665 patients undergoing hip surgery, HIT occurred in 9 of 332 patients receiving unfractionated heparin (2.7 percent) and none of those receiving LMW heparin [42]. Eight of the nine patients with HIT had a thrombotic event (seven venous, one arterial), whereas 117 of the patients who did not develop HIT had a VTE (89 versus 18 percent, respectively).

●In a series of 260 patients with HIT, upper extremity venous thrombosis occurred exclusively in patients with a central venous catheter (14 of 145 with a central venous catheter [10 percent]), always in association with the catheter [68].

Thrombotic sequelae — Complications of thrombosis include death (most commonly due to pulmonary embolism), skin necrosis, limb gangrene (sometimes requiring amputation), and organ infarction.

●Skin necrosis – Skin necrosis at the site of heparin injections (picture 1 and picture 2) should immediately suggest HIT [69-72]. Some patients with HIT have skin necrosis without thrombocytopenia [70,71,73]. Skin findings in locations other than the injection site have also been reported [74]. Affected areas are usually in fat-rich areas such as the abdomen; however, the distal extremities and the nose can also be involved. The appearance of erythema is followed by purpura and hemorrhage with subsequent necrosis.

●Limb gangrene – Limb gangrene was associated with venous rather than arterial thrombosis in 8 of 158 patients with HIT who developed this complication [75]. In one study, patients with venous limb gangrene had a much higher international normalized ratio (INR) than patients who did not develop gangrene (median INR: 5.8 versus 3.1), suggesting a possible contribution from acquired protein C deficiency and providing a rationale for stopping and reversing warfarin in all patients with suspected HIT who have ongoing thrombocytopenia [75]. (See "Protein C deficiency", section on 'Warfarin-induced skin necrosis'.)

●Organ ischemia or infarction – Arterial thrombosis can lead to stroke, myocardial infarction, acute limb ischemia from peripheral arterial occlusion, or organ infarction (mesentery, kidney) [76,77]. Since the arterial circulation is a high-flow, high shear-rate environment, thrombi tend to be platelet-rich (ie, "white clots"). Other unusual thrombotic sequelae include adrenal insufficiency secondary to adrenal vein thrombosis, and transient global amnesia, possibly due to brain ischemia [78-80].

Anaphylaxis — Acute systemic anaphylactic reactions have been described in patients with HIT, and these can be fatal [81]. Thrombocytopenia may be absent despite clinical and laboratory evidence of HIT [72].

EVALUATION

Suspecting HIT — Any one of the following scenarios should raise the possibility of HIT in patients who are currently receiving heparin or who received heparin in the preceding 5 to 10 days [82,83]:

●New onset of thrombocytopenia (ie, platelet count <150,000/microL)

●A decrease in platelet count by 50 percent or more, even if the platelet count exceeds 150,000/microL

●Venous or arterial thrombosis

●Necrotic skin lesions at heparin injection sites

●Acute systemic reactions (eg, fever/chills, tachycardia, hypertension, dyspnea, cardiopulmonary arrest) occurring after intravenous heparin administration

Importantly, one should not wait for thrombosis to develop before suspecting HIT, because thrombocytopenia often precedes thrombosis. Early intervention has the potential to prevent thrombotic events, which are the major cause of morbidity and mortality in patients with HIT.

Overview of our approach to evaluation — We always consider clinical and laboratory evidence in evaluating patients for HIT (algorithm 1). However, definitive laboratory data (ie, immunoassay and/or functional assay for HIT antibodies) may not be available for several days. Thus, we make a presumptive diagnosis of HIT based on clinical findings and laboratory data that are immediately available (eg, platelet count, same-day HIT testing), and we confirm or refute the diagnosis once we have the definitive results of HIT antibody testing.

The 4 Ts score (calculator 1) is an easy-to-use score that quantifies the clinical findings associated with HIT and helps to establish the pretest probability (likelihood) of HIT (table 4). The 4 Ts score is presented below and in tabular form from the American Society of Hematology (ASH). A 2018 clinical practice guideline from ASH recommends use of the 4 Ts score rather than "clinical gestalt" to assess the pretest probability of HIT [2]. (See '4 Ts score' below.)

The HIT Expert Probability (HEP) score involves a more detailed assessment and assigns positive and negative points for a number of different clinical features and alternative causes of thrombocytopenia [84]. (See 'Other pretest probability scores' below.)

●If the 4 Ts score is intermediate or high probability, we make a presumptive diagnosis of HIT pending results of HIT antibody testing. Such scores will capture the vast majority of patients with HIT.

●If the 4 Ts score is low probability, we generally do not pursue HIT antibody testing or presumptive treatment for HIT (eg, we do not discontinue heparin or start a non-heparin anticoagulant) because the risk of HIT is low and presumptive treatment carries risks (eg, bleeding) and costs. In rare or complex cases, the clinician may suspect HIT in an individual with a low probability score (eg, platelet count just outside cutoff).

Any change in clinical findings (eg, development of a new thrombosis, further drop in platelet count) should prompt re-evaluation with HIT antibody testing if appropriate.

The 4 Ts score should be used as a guide for clinicians and should not substitute for clinical judgment. The scoring system has not been compared with the accuracy of intuition-based diagnosis. Thus, clinicians, particularly those who have less familiarity with HIT, should consult someone more knowledgeable in this diagnosis because a presumptive diagnosis of HIT carries management implications (eg, discontinuation of heparin, administration of a non-heparin anticoagulant). Not diagnosing HIT when it is present or mistakenly diagnosing HIT when it is absent are both associated with significant risks (eg, life-threatening thrombosis if a diagnosis is missed; life-threatening bleeding and high cost of presumptive treatment with a non-heparin anticoagulant if an incorrect diagnosis is made) [85].

A presumptive diagnosis of HIT should not be delayed while awaiting results of HIT antibody testing. However, if results of HIT antibody testing are immediately available, these may be helpful, as discussed below. (See 'Approach to HIT antibody testing' below.)

Individuals with suspected HIT should have immediate discontinuation of all sources of heparin, and administration of a non-heparin anticoagulant unless there is bleeding or a high risk of bleeding (algorithm 1). (See "Management of heparin-induced thrombocytopenia".)

All patients with a presumptive diagnosis of HIT should have laboratory testing for HIT antibodies. This testing is challenging because HIT immunoassays (eg, enzyme-linked immunosorbent assay [ELISA] for anti-platelet factor 4 [PF4] antibodies) are readily available in many centers but have low specificity, whereas a functional assay such as a serotonin release assay (SRA), which measures the ability of patient serum to activate test platelets in the presence of heparin, is definitive in the vast majority of cases but may take several days to return. Our approach to the use of this testing is described below. (See 'Approach to HIT antibody testing' below.)

Individuals with a confirmed diagnosis of HIT should continue the non-heparin anticoagulant. (See "Management of heparin-induced thrombocytopenia".)

If the diagnosis of HIT is excluded, heparin(s) can be used if/when indicated. We continue to monitor the patient clinically and pursue other causes of thrombocytopenia. (See 'Differential diagnosis' below and "Diagnostic approach to thrombocytopenia in adults".)

This approach is consistent with the 2018 ASH practice guideline on HIT [2].

Approaches to evaluating thrombocytopenia in children and neonates and reviews of other potential causes of thrombocytopenia in children and neonates are presented separately. (See "Approach to the child with unexplained thrombocytopenia" and "Neonatal thrombocytopenia: Clinical manifestations, evaluation, and management" and "Causes of thrombocytopenia in children" and "Neonatal thrombocytopenia: Etiology".)

Clinical findings

4 Ts score

Calculating the score — The 4 Ts score is used for estimating the likelihood (pretest probability) of HIT based on readily available clinical features (calculator 1) [86]. It is used to make a presumptive diagnosis of HIT until laboratory data are available, and is then integrated with laboratory data to make a final diagnosis.

The score assesses the degree of thrombocytopenia, the timing relative to heparin exposure, the presence of thrombosis, and other causes for thrombocytopenia. Points are assigned as follows (table 4):

●Thrombocytopenia

•Platelet count fall >50 percent and nadir ≥20,000/microL – 2 points

•Platelet count fall 30 to 50 percent or nadir 10 to 19,000/microL – 1 point

•Platelet count fall <30 percent or nadir <10,000/microL – 0 points

●Timing of platelet count fall

•Clear onset between days 5 and 10 or platelet count fall at ≤1 day if prior heparin exposure within the last 30 days – 2 points

•Consistent with fall at 5 to 10 days but unclear (eg, missing platelet counts), onset after day 10, or fall ≤1 day with prior heparin exposure within 30 to 100 days – 1 point

•Platelet count fall at <4 days without recent exposure – 0 points

●Thrombosis or other sequelae

•Confirmed new thrombosis, skin necrosis, or acute systemic reaction after intravenous unfractionated heparin bolus – 2 points

•Progressive or recurrent thrombosis, non-necrotizing (erythematous) skin lesions, or suspected thrombosis that has not been proven – 1 point

•None – 0 points

●Other causes for thrombocytopenia

•None apparent – 2 points

•Possible – 1 point

•Definite – 0 points

Interpretation — The sum of the point values gives a total from 0 to 8. Pretest probabilities for HIT are as follows (table 5) [87]:

●0 to 3 points – Low probability (risk of HIT <1 percent)

●4 to 5 points – Intermediate probability (risk of HIT approximately 10 percent)

●6 to 8 points – High probability (risk of HIT approximately 50 percent)

The 4 Ts score captures the major clinical features of HIT and the likelihood that these findings are due to heparin rather than another cause. Patients receiving heparin have a higher-than-average baseline risk of thrombosis; thus, the isolated development of venous thromboembolism (VTE) or myocardial infarction alone will not generate an intermediate or high probability 4 Ts score. The 4 Ts score has not been validated for patients receiving low molecular weight (LMW) heparin, although we use it in this population.

Supporting evidence for limiting laboratory testing for HIT to patients with an intermediate or high probability 4 Ts score or in complex clinical settings in which HIT is suspected clinically but the 4 Ts score is less reliable because of missing or incomplete information include [88-91]:

●In a prospective study of 111 patients with a low probability of HIT, only one had clinically significant HIT antibodies (0.9 percent) [86]. Among those with intermediate and high probability 4T scores, clinically significant HIT antibodies were present in 11.4 and 34 percent, respectively.

●In a meta-analysis of 3068 patients with clinically suspected HIT, the proportion of individuals with a low probability 4 Ts score who had a negative functional assay for HIT antibodies (negative predictive value [NPV]) was 0.998 (95% CI 0.97-1.0) [90]. The proportions of those with an intermediate or high probability score who had a positive functional assay for HIT antibodies (positive predictive values [PPVs]) were 0.14 and 0.64, respectively.

●In a study of 526 patients with possible HIT, 6 of 321 with a low probability 4 Ts score (1.9 percent) had a positive functional assay (SRA), resulting in the NPV for a low probability 4 Ts score of 98.1 percent [92]. Addition of a rapid gel immunoassay (PF4/H-PaGIA) improved the NPV to 100 percent. However, such a gel immunoassay with rapid turnaround is not widely available.

The above studies primarily (or exclusively) included adults, since HIT is very rare in children (see 'Incidence and risk factors' above). In a 2022 study that compared 4Ts scores in a cohort of 34 children and 105 adults evaluated for HIT in the same institution over the same time period, the children overall had higher 4Ts scores and lower rates of confirmed HIT in a functional assay (16 of 105 adults [15 percent] tested positive, versus 3 of 34 children [9 percent]) [57]. All three of the children with HIT presented with thrombosis.

Other pretest probability scores — Two other pretest probability scores have been proposed:

●Lillo-Le Louët model – This model, which is intended for use exclusively in the postcardiopulmonary bypass setting, awards points based on the timing and duration of thrombocytopenia and the duration of cardiopulmonary bypass [93]. This model has not been prospectively evaluated on a multicenter basis.

●HIT Expert Probability (HEP) Score – This score was developed based on broad expert opinion [84]. Points are awarded for the timing and degree of thrombocytopenia, with finer resolution than the 4 Ts score, and points are subtracted for bleeding and other causes of thrombocytopenia. In a comparison study, the experts who developed this score found it to have greater interobserver agreement than the 4 Ts score. In a prospective cohort of 310 patients with suspected HIT who had simultaneous calculation of the 4 Ts score and the HEP score before the results of HIT antibody testing were known, the HEP score provided similar diagnostic accuracy overall [94]. However, the HEP score had better performance characteristics (higher sensitivity and specificity) for determining the likelihood of HIT in patients in the intensive care unit and when used by clinicians who had less experience (eg, trainees). However, a prospective study of 51 patients with suspected HIT that compared the HEP score with the 4Ts score found no significant differences between the scores compared with the diagnostic standard (ie, positive SRA) [95].

We do not use either of these models routinely, although others may do so.

HIT antibody testing

Approach to HIT antibody testing — The diagnosis of HIT is based on clinical features supported by laboratory testing. Laboratory testing includes two major types of tests for anti-PF4-heparin antibodies (immunoassays and functional assays):

●Immunoassays – Immunoassays detect the presence of a PF4-heparin antibody but not its ability to bind and activate platelets (see 'Immunoassay (eg, ELISA)' below). The results are reported in optical density (OD) units. These assays are widely available, fast, and straightforward to interpret, but they have a higher incidence of false positive results than functional assays; they may also detect antibodies that are not clinically significant. While the assays can be performed rapidly, rapid turnaround of results cannot be assumed, as some laboratories batch assays and only perform them on certain days of the week. Clinicians should speak with the laboratory if there is increased urgency in obtaining results. Clinicians should also be aware that there is variability among different immunoassays. The ranges of sensitivities, specificities, and predictive values of various commercially available immunoassays and rapid immunoassays were described in two systematic reviews and meta-analyses that included data from over 15,000 patients [96,97].

●Functional assays – Functional assays assess the ability of a PF4-heparin antibody to bind and activate platelets, and thus to cause the clinical HIT syndrome (see 'Functional assays' below). The results are reported as positive or negative. Functional assays are more resource-intensive and require more expertise to perform and interpret, but they are considered more specific for confirming a diagnosis of HIT and may be especially helpful if the immunoassay gives an indeterminate result [98]. Most institutions do not have an in-house functional assay, but these assays are commonly available as a send-out test. Thus, the turnaround time is usually longer than that of an ELISA (eg, days).

One exception to the use of a functional assay is a patient receiving the anti-platelet drug ticagrelor. (See 'Patients receiving ticagrelor' below.)

The likelihood of HIT is greater with a higher pretest probability (eg, 4 Ts score) and with a higher OD on the PF4-heparin ELISA. However, many individuals with an intermediate or even a high probability 4 Ts score, as well as many individuals with a positive ELISA, will not have HIT. In addition, ELISA OD units may vary among clinical laboratories depending on the heparin-PF4 ELISA kit used. As a result, there is no universally agreed-upon OD value that can be labeled as "positive," and different experts use different OD cutoffs [82,87,99,100]. One should consult with the local institutional laboratory regarding its OD threshold.

We perform HIT antibody testing as follows (algorithm 1):

●For the vast majority of individuals with a low probability 4 Ts score, we do not obtain laboratory testing for HIT. The likelihood of HIT in these individuals is <1 percent [87]. If the pretest probability of HIT changes over time, it is reasonable to repeat a previously negative ELISA test because a positive ELISA test under such circumstances is more likely to be associated with clinical HIT [101].

●For all patients with an intermediate or high probability 4 Ts score, we obtain a PF4-heparin ELISA (immunoassay). We continue to treat for a presumptive diagnosis of HIT until definitive results of the ELISA and/or the functional assay are available (table 5).

•For any 4 Ts score, if the ELISA OD is <0.60, we consider the diagnosis of HIT to be ruled out. Rarely, we may obtain a functional assay if the clinical picture and ELISA are discordant (eg, high probability 4 Ts score and ELISA OD <0.40).

•For patients with an intermediate probability 4 Ts score and an OD ≥2.00, we consider the diagnosis of HIT to be confirmed; if the OD is between 0.60 and 1.99, we obtain a functional assay.

•For patients with a high probability 4 Ts score and an OD ≥1.50, we consider the diagnosis of HIT to be confirmed; if the OD is between 0.60 and 1.49, we obtain a functional assay.

•We consider the diagnosis of HIT to be confirmed if the functional assay is positive and we consider the diagnosis of HIT excluded if the functional assay is negative, although there are rare individuals with a disorder clinically consistent with HIT who have a negative functional assay [102]. Different functional assays at different institutions may have different performance characteristics, and clinicians should become familiar with the performance characteristics of the functional assay used at their institution.

•If a functional assay is not readily available, we use an OD >1.00 as confirmation of HIT.

●Some centers may favor the use of functional assays in all patients, even those with an ELISA OD ≥2.00.

Useful data analyses and decision support for determining the need for a functional assay are described in the following studies:

●A 2013 study analyzed data from 1958 patients suspected of having HIT, used a combination of 4 Ts scores and ELISA assay receiver operator curves, conducted a Bayesian analysis, and produced a decision-support table [87].

●Another Bayesian analysis estimated the post-test probability of HIT for a variety of combinations of 4 Ts scores and ELISA OD results and came to similar conclusions regarding appropriate settings for a functional assay [99].

●A more generalizable tool for using the 4 Ts score and a "positive" or "negative" immunoassay result to determine the likelihood of HIT has been developed [100]. However, it may be important for clinicians to familiarize themselves with the ELISA assay used at their institution to fully evaluate and interpret a "positive" result.

Immunoassay (eg, ELISA) — The solid-phase ELISA is the most widely used laboratory test for HIT. This is an immunoassay that detects the presence of anti-platelet factor 4 (PF4)/heparin antibodies in patient serum. The turnaround time for ELISA testing varies among centers, often as a reflection of the volume of HIT testing performed, which determines how frequently a batch of tests is run. As noted below, rapid immunoassays that can be performed on single samples (turnaround time 20 to 30 minutes) are being developed and some appear to have comparable accuracy to the ELISA.

The assay is performed by adding patient serum to a microtiter plate coated with heparin-PF4 complexes or polyanion-PF4 complexes that generate the HIT antigen on PF4. If HIT antibodies are present in the patient sample, they will bind the PF4 complex and can be identified by adding a second antibody coupled to an enzyme for detection. The intensity of a colorimetric change correlates with the concentration of the HIT antibody in the patient sample. This intensity is reported as the optical density (OD) at a specific wavelength (typically 405 nm). A higher OD represents a higher titer of antibody in the patient's serum, which is more strongly suggestive of HIT than a low titer.

Most laboratories report the results of a HIT ELISA assay as positive or negative based on a predetermined OD cutoff value. However, it is important that laboratories make available the OD of the ELISA assay because the OD correlates with the likelihood of HIT [88,103-105]. If the OD is not reported, it should be obtained from the laboratory.

A study of 1553 patient samples tested the correlation between the OD in a HIT ELISA and a functional laboratory test considered to be the gold standard (ie, a serotonin release assay [SRA]) [104]. (See 'Serotonin release assay' below.)

In this study, the likelihood of a positive SRA was as follows [104]:

●OD <0.40 – SRA positive in 0.0 to 0.1 percent

●OD 0.40 to <1.00 – SRA positive in 1 to 5 percent

●OD 1.00 to <1.40 – SRA positive in 18 to 30 percent

●OD 1.40 to <2.00 – SRA positive in 50 to 58 percent

●OD >2.00 – SRA positive in 89 to 100 percent

The authors concluded that a weakly positive ELISA (ie, OD 0.40 to <1.00) was strong evidence against a diagnosis of HIT, with a probability of <5 percent. At an OD ≥1.40, the probability of HIT became more likely than not, and at an OD >2.00, the probability of HIT was 90 percent or greater [104]. As noted above, however, other assays have used different thresholds (eg, OD <0.60 as criteria to determine that HIT is not present (table 5)), and the threshold ODs to use as evidence against or for HIT, or that a functional assay is not needed may vary with different ELISA assays. (See 'Approach to HIT antibody testing' above.)

Another study evaluated the correlation between the OD of a HIT immunoassay with thrombosis in 318 consecutive hospitalized patients with clinically suspected HIT [106]. In this study, the overall rate of arterial or venous thrombosis was 23.3 percent, and the 30-day mortality was 16.7 percent. For every 1-unit increase in the OD, there was an approximate doubling in the odds of thrombosis (odds ratio [OR] 1.9; 95% CI 1.5-2.6). Regression analysis predicted that the proportion of patients with clinically evident thrombosis at an OD of 1.0, 2.0, and 3.0 were 19, 32, and 48 percent, respectively.

The high sensitivity of the ELISA (ie, a negative test strongly suggests HIT is not the diagnosis) has been corroborated by other studies, with reported sensitivities of 91 to over 97 percent. Thus, if the OD is <0.60, we do not perform further testing unless the clinical findings change (table 5), because the likelihood of HIT is exceedingly low. The low specificity of the ELISA (ie, a positive test may occur in the setting of non-HIT conditions) has also been noted in various studies [107-109].

A rapid liquid phase immunoassay has also been developed with a high negative predictive value (NPV; ie, a high proportion of negative tests that are true negatives). In a cohort of 334 consecutive patients, the sensitivity and specificity of this assay were 97.5 and 83.4 percent, respectively, with an NPV of 99.6 percent and a positive predictive value (PPV) of 44.4 percent, relative to an SRA [110]. A meta-analysis of studies comparing rapid immunoassays found that all of these assays had high sensitivity (0.96 to 1.00), with wider variability in specificity (0.68 to 0.94) [97]. In general, the sensitivity of these assays was as good as solid phase ELISA testing, and the specificity was better, although one rapid immunoassay (the PIFA) did not perform as well.

Another result that can be reported from the immunoassay is the ability of higher doses of heparin to titrate away the HIT antibody and reduce the OD. This phenomenon is based on the binding of HIT antibodies to PF4-heparin over a narrow range of heparin concentrations. In a study of 115 patients whose samples were tested in the presence of high- and low-dose heparin, the OD of the original ELISA using low-dose heparin and the results of high-dose heparin testing to confirm HIT antibodies (ie, to cause a >50 percent reduction in the OD) was used to create a nomogram for predicting the probability of HIT [105].

The specificity of immunoassays can be further improved by assaying only for IgG antibodies to PF4/heparin, rather than IgM, IgG, and IgA, since IgM and IgA antibodies were shown to be less specific for diagnosing HIT in many studies [39,89,111,112].

Functional assays — Functional assays test the ability of HIT antibodies from patient serum to activate test platelets, which more strongly correlates with the presence of HIT than the results of an immunoassay. We reserve functional assays for those with indeterminate immunoassays or those for whom the clinical picture and immunoassay results are discordant (eg, patient with very high probability 4 Ts score and negative immunoassay). In centers where a functional assay is available rapidly, it may be appropriate to proceed directly to a functional assay rather than start with an immunoassay.

Serotonin release assay — The SRA measures platelet activation by detecting the release of serotonin from test platelets in the presence of patient serum and heparin. Either unfractionated heparin or low molecular weight heparin can be used in the assay; the standard assay uses unfractionated heparin.

Test platelets from normal donors are radiolabeled with ¹⁴C-serotonin and incubated with patient serum plus heparin at therapeutic or excessive concentrations. A positive test is the release of ¹⁴C-serotonin when therapeutic heparin concentrations are used (0.1 units/mL), but not with much higher heparin concentrations (100 units/mL). This concentration dependence reflects the fact that the binding of HIT antibodies is only seen at certain ratios of heparin to PF4. (See 'Formation of HIT antibodies' above.)

The SRA is considered the gold standard among diagnostic tests for HIT, although this has been brought into question by some experts [113]. It has a sensitivity and specificity of >95 percent when performed by experienced laboratories [88,114]. The disadvantages of this test are its high cost, the use of radioactive material, the technical demands of the assay, and the delay in obtaining results due to lack of routine availability at most institutions. There are also institution-to-institution differences in the techniques used in the test that may cause differing performance characteristics (eg, whether washed platelets are used).

A nonradioactive assay in which serotonin release is quantitated via a commercially available ELISA has been described, with a sensitivity and specificity of 100 and 97 percent, respectively [115]. However, this assay is not generally available.

Heparin-induced platelet activation (HIPA) — In the heparin-induced platelet activation (HIPA) assay, serum or platelet-poor plasma from a patient with suspected HIT is added to platelet-rich plasma (PRP) from healthy donors, and platelet activation (generally measured by aggregation) is assayed in the absence of heparin and in the presence of low and high heparin concentrations. The use of washed platelets increases the sensitivity of the test. A positive test shows minimal platelet activation in the absence of heparin and in the presence of high heparin concentrations (eg, 10 to 100 units/mL) and strong activation in the presence of low heparin concentrations (eg, 0.1 to 0.3 units/mL).

The HIPA assay is the more widely used of the functional assays, especially in Europe [116].

Assays under development

PF4-dependent P-selectin expression — The PF4-dependent P-selectin expression assay (PEA) is a functional HIT assay that does not require radioactive reagents. Unlike the SRA, the PEA involves incubation of patient serum with washed test platelets pre-treated with PF4 (rather than heparin). The PEA measures expression of P-selectin on the platelet surface rather than release of radiolabeled ¹⁴C-serotonin to detect platelet activation by HIT antibodies [117]. In a 2016 study that compared the results of the PEA and SRA assays using banked serum from 91 patients who had results for the 4 Ts score available, the PEA had greater accuracy for HIT than the SRA (area under the receiver operator curve: 0.92 versus 0.82). However, it is not clear whether the 16 individuals with a negative SRA and a positive PEA in fact had clinically important HIT [117,118].

DIAGNOSIS — HIT is diagnosed by integrating clinical features and laboratory testing; neither of these alone is sufficient (algorithm 1).

●A presumptive diagnosis of HIT must often be made purely on clinical findings and platelet counts (eg, intermediate to high probability 4 Ts score) until the results of HIT antibody testing are available. (See '4 Ts score' above.)

●In individuals with a presumptive diagnosis of HIT based on the 4 Ts score, the diagnosis is considered to be confirmed if there is a positive enzyme-linked immunosorbent assay (ELISA) with an optical density (OD) ≥2.00 (or ≥1.50 for patients with a high probability 4 Ts score) or if there is a positive functional assay. (See 'HIT antibody testing' above.)

●Additional considerations may apply to individuals with HIT variants in which there was no heparin exposure. (See 'Individuals not exposed to heparin' below.)

DIAGNOSIS IN SPECIAL POPULATIONS — Evaluation and diagnosis of HIT can be more complicated in some patient populations due to extensive use of heparin and a high incidence of heparin-induced antibodies without clinical HIT.

Individuals not exposed to heparin — Heparin exposure is not always obvious, as it may not be noted in medication lists or anesthesia records. However, there are rare individuals who present with a HIT-like syndrome who are determined after a thorough evaluation not to have been exposed to any heparin, and there may be a concern for HIT variants such as spontaneous HIT or vaccine-induced immune thrombotic thrombocytopenia (VITT). (See 'Terminology and HIT variants' above.)

●Spontaneous HIT – Spontaneous HIT is clinically indistinguishable from HIT except for the lack of prior heparin exposure. (See 'Terminology and HIT variants' above.)

Evaluation for spontaneous HIT is identical to HIT, including laboratory testing and interpretation. (See 'HIT antibody testing' above.)

●VITT – VITT has been reported after receipt of certain adenoviral-vectored COVID-19 vaccines. Diagnosis of VITT is discussed separately. (See "COVID-19: Vaccine-induced immune thrombotic thrombocytopenia (VITT)", section on 'Evaluation'.)

Children — There are no distinct validated diagnostic paradigms for children. In the absence of pediatric-specific algorithms, evaluation is similar to adults. (See 'Evaluation' above.)

Dialysis — Patients undergoing continuous renal replacement therapy or intermittent hemodialysis typically require anticoagulation to prevent clotting of the extracorporeal circuits, and heparin is most commonly used. Chronic heparin exposure is associated with a high incidence of heparin-induced antibodies, the clinical significance of which is uncertain.

We do not routinely test patients undergoing dialysis for HIT antibodies, although we do evaluate and treat these patients for HIT in the appropriate clinical setting. A high incidence of clinically insignificant heparin-induced antibody formation was observed in a group of 740 patients with renal failure upon initiation of hemodialysis or peritoneal dialysis [119]. The incidence of heparin-induced antibodies by immunoassay was highest during the first 90 days of hemodialysis (20 percent). By six months, approximately 9 percent of patients undergoing hemodialysis or peritoneal dialysis had antibodies. A positive immunoassay did not correlate with thrombocytopenia or thrombosis, although nine patients did develop clinical HIT during the study (1.2 percent).

The management of patients with a history of HIT or suspected HIT undergoing continuous or intermittent hemodialysis is discussed separately. (See "Anticoagulation for continuous kidney replacement therapy" and "Anticoagulation for the hemodialysis procedure", section on 'Patients with heparin-induced thrombocytopenia'.)

Cardiac surgery/cardiopulmonary bypass — The diagnosis of HIT in patients undergoing cardiopulmonary bypass (CPB) is challenging because these patients receive large amounts of heparin and normally have a decrease in the platelet count of approximately 40 to 50 percent during the first 72 hours following surgery. Approaches to evaluation and management of acute HIT in patients undergoing cardiac or vascular surgery, as well as performance of cardiac or vascular surgery in patients with a prior history of HIT, are discussed separately. (See "Management of heparin-induced thrombocytopenia (HIT) during cardiac or vascular surgery".)

Patients receiving ticagrelor — The product information for the antiplatelet agent ticagrelor was updated to state that may interfere with functional assays for HIT, making it appear that the test is negative when in fact HIT is present [120,121]. This proposed effect was inferred by inhibition of the P2Y₁₂ receptor on normal platelets used in the assay. Whether other antiplatelet agents or combinations of agents will cause similar changes is unknown.

An alternative functional assay should be performed, or, if one is not readily available, the 4 Ts score and the HIT immunoassay results (particularly its OD value) should be used to guide recommendations and therapy.

DIFFERENTIAL DIAGNOSIS — Differentiation between HIT and other causes of thrombocytopenia is important because a diagnosis of HIT requires use of a non-heparin anticoagulant that is likely to be expensive and may be associated with bleeding and lifelong heparin avoidance. The distinction between HIT and other causes of thrombocytopenia can be difficult because many medical conditions and medications can cause thrombocytopenia, and heparin use in hospitalized patients is common, sometimes without being obvious (eg, as heparin flushes or in the operating room) [83].

The differential diagnosis of HIT is large and includes the following:

●Disseminated intravascular coagulation, sepsis, or infection – Thrombocytopenia is a common complication of severe infections, due to platelet consumption in disseminated intravascular coagulation (DIC) or bone marrow suppression from infection. (See "Evaluation and management of disseminated intravascular coagulation (DIC) in adults", section on 'Differential diagnosis'.)

•HIT and DIC can both be accompanied by thrombosis, although patients with DIC may also have bleeding (eg, oozing from catheter sites).

•Unlike most cases of HIT, DIC is associated with abnormal coagulation studies (eg, prolonged prothrombin time [PT] and activated partial thromboplastin time [aPTT], low fibrinogen, elevated D-dimer). Severe cases of HIT can also be associated with a consumptive coagulopathy [4]. Unlike HIT, infection-associated thrombocytopenia is often accompanied by other signs of bone marrow suppression (eg, anemia, leukopenia, decreased bone marrow megakaryocytes).

●Immune thrombocytopenia – Immune thrombocytopenia (ITP, previously called immune thrombocytopenic purpura) is a common autoimmune disorder. (See "Immune thrombocytopenia (ITP) in adults: Clinical manifestations and diagnosis".)

•ITP and HIT are both characterized by thrombocytopenia and increased bleeding risk.

•Unlike HIT, ITP lacks a clear temporal relationship to heparin exposure, and ITP is not associated with thrombosis. Laboratory testing for antiplatelet antibodies is used to diagnose HIT, but not for patients thought to have ITP. Platelet counts are widely variable in patients with ITP; a very low platelet count (<20,000/microL) is unusual in HIT but common in ITP. HIT resolves upon withdrawal of heparin, while ITP persists for months or indefinitely.

●Post-transfusion purpura – Post-transfusion purpura (PTP) is a very rare antibody-mediated transfusion reaction that occurs after transfusion of red blood cells or platelets, or in an individual previously sensitized to platelet antigens during pregnancy. (See "Immunologic transfusion reactions", section on 'Post-transfusion purpura'.)

•HIT and PTP are both caused by antibodies, but only HIT antibodies activate platelets and can cause thrombosis.

•The clinical history and laboratory testing differ for the two conditions (heparin exposure and HIT antibody testing for HIT; prior pregnancy or recent transfusion, with human platelet antigen 1a for PTP). In those who have had a recent transfusion and exposure to heparin, distinction between HIT and PTP may depend on the presence of thrombosis or results of laboratory testing.

●Thrombotic microangiopathy – Thrombotic microangiopathies (TMAs) are a group of disorders associated with mechanical red blood cell fragmentation and thrombocytopenia due to vascular lesions that trap red cells and platelets. These include thrombotic thrombocytopenic purpura (TTP), hemolytic uremic syndrome (HUS), and other TMAs associated with chemotherapeutic agents or renal transplantation. (See "Diagnostic approach to suspected TTP, HUS, or other thrombotic microangiopathy (TMA)" and "Diagnosis of immune TTP".)

•HIT and TMAs are both characterized by thrombocytopenia and thrombosis.

•Unlike HIT, TMAs are characterized by many fragmented red blood cells (ie, schistocytes) on the peripheral blood smear and microvascular rather than large vessel thrombosis. Laboratory testing for TMAs also differs from HIT, as described separately.

●Drug-induced thrombocytopenia – Medications and certain foods can cause thrombocytopenia by inducing drug-dependent, platelet reactive antibodies; common examples include antibiotics, glycoprotein IIb/IIIa inhibitors, and quinine. Protamine sulfate, used to reverse the anticoagulant effect of heparin, has also been reported to cause thrombocytopenia. (See "Drug-induced immune thrombocytopenia".)

•Both conditions are associated with thrombocytopenia, often in a hospitalized patient.

•Unlike HIT, thrombocytopenia induced by medications or food usually is not associated with thrombosis (with rare exceptions such as drugs that cause a TMA). In drug-induced thrombocytopenia, discontinuation of the drug may be sufficient treatment, whereas in HIT, anticoagulation is required as well.

●Venous thromboembolism unrelated to heparin – Hospitalized patients are at risk for venous thromboembolism (VTE; ie, deep vein thrombosis, pulmonary embolism), may develop multifactorial thrombocytopenia, and often receive heparin for thromboprophylaxis. HIT can be associated with VTE or arterial emboli, whereas hospitalized patients may develop VTE as a consequence of their hospital admission but generally do not develop arterial emboli. The differentiation of HIT from "hospital-associated" VTE may be difficult, particularly in critically ill patients. (See "Overview of the causes of venous thrombosis".)

●Systemic lupus erythematosus and/or antiphospholipid syndrome – Mild thrombocytopenia is common in patients with systemic lupus erythematosus (SLE; eg, platelet count between 50,000 and 150,000/microL). (See "Hematologic manifestations of systemic lupus erythematosus", section on 'Thrombocytopenia' and "Clinical manifestations of antiphospholipid syndrome", section on 'Hematologic abnormalities'.)

•HIT and SLE-associated thrombocytopenia are both caused by antibodies; HIT antibodies activate platelets and can cause thrombosis, whereas most SLE-associated antiplatelet antibodies do not. An exception is antiphospholipid antibodies in individuals with antiphospholipid syndrome (APS), which may occur in the setting of SLE or independently. APS is commonly accompanied by mild thrombocytopenia. Like HIT, APS can be accompanied by venous and arterial thrombosis.

•Unlike HIT antibodies, APS antibodies are not heparin-dependent. Unlike HIT, APS is characterized by specific antiphospholipid antibodies.

●Delayed-type hypersensitivity and allergic skin reactions – Administration of unfractionated or low molecular weight (LMW) heparin can cause skin reactions at the injection site due to delayed-type hypersensitivity (DTH). These are typically eczematous or pruritic erythematous plaques without necrosis. (See "Heparin and LMW heparin: Dosing and adverse effects", section on 'Skin necrosis and local allergic reactions'.)

Unlike HIT, these DTH reactions are not associated with thrombosis. Skin biopsy, if done, shows dermal microvascular thrombi in HIT, and lymphohistiocytic perivascular dermal infiltrate, sometimes with eosinophils, in DTH. Laboratory testing (eg, enzyme-linked immunosorbent assay [ELISA] testing for platelet factor 4 [PF4]) can also be done to distinguish these conditions. Switching between heparin products may eliminate these reactions.

●Other causes of thrombocytopenia – A variety of other causes of thrombocytopenia (eg, myelodysplasia, HIV infection) may coincidentally occur in patients exposed to heparin (table 6). These are discussed separately. (See "Diagnostic approach to thrombocytopenia in adults" and "Causes of thrombocytopenia in children".)

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" and "Society guideline links: Heparin-induced thrombocytopenia (HIT)".)

SUMMARY AND RECOMMENDATIONS

●Pathophysiology – Heparin-induced thrombocytopenia (HIT) is a life-threatening complication of exposure to unfractionated or low molecular weight (LMW) heparin (prevalence, up to 5 percent). HIT is caused by autoantibodies to platelet factor 4 (PF4) complexed with heparin that activate platelets (figure 1). Rare HIT variants have been reported in the absence of heparin (table 3). (See 'Terminology and HIT variants' above and 'Pathophysiology' above.)

●Risk factors – Risk factors include surgery, unfractionated heparin, higher heparin doses, female sex, and possibly age. (See 'Incidence and risk factors' above.)

●Presentation – All patients have a platelet count decrease; thrombocytopenia (platelets <150,000/microL) occurs in 90 percent. Thrombosis occurs in up to 50 percent of untreated patients, venous more common than arterial. Thrombosis can cause skin necrosis, limb gangrene, and organ infarction. An anaphylaxis-like reaction to heparin can occur. (See 'Clinical manifestations' above.)

●Evaluation – We always consider clinical and laboratory evidence (algorithm 1). Laboratory results may not be available for several days, and it is often necessary to make a presumptive diagnosis. Failing to diagnose HIT when present and mistakenly diagnosing HIT when absent carry significant risks (life-threatening thrombosis or bleeding, respectively). (See 'Overview of our approach to evaluation' above.)

●4 Ts score – The 4 Ts score (calculator 1) estimates the likelihood of HIT based on the degree of thrombocytopenia, timing of platelet count drop, presence of thrombosis, and absence of other causes of thrombocytopenia. (See '4 Ts score' above.)

•Intermediate or high probability – Presumptively diagnose HIT (table 4). These patients should immediately discontinue all heparins and receive a non-heparin anticoagulant to reduce the risk of thrombosis, unless there is active (or a high risk of) bleeding. (See 'Diagnosis' above.)

•Low probability – Risk of HIT is very low (table 4). Generally do not pursue further testing; however, in complex cases, HIT may be present (platelet count just outside cutoff, uncertainty about alternative causes of thrombocytopenia) and may warrant a presumptive diagnosis of HIT.

●Laboratory testing – We confirm or refute the diagnosis with HIT antibody testing. We generally start with an immunoassay (enzyme-linked immunosorbent assay [ELISA], rapid immunoassay). (See 'HIT antibody testing' above.)

Interpretation is as follows (table 5):

•OD <0.60 – HIT excluded. Omit further HIT testing unless the clinical picture changes or is discordant (high probability 4 Ts score).

•OD 0.60 to 1.99 (0.6 to 1.49 for high probability 4 Ts score) – Indeterminate. Continue non-heparin anticoagulant and obtain a functional assay.

•OD ≥2.00 (≥1.50 for high probability 4 Ts score) – HIT confirmed. Functional HIT antibody testing not required.

Positive functional assays confirm HIT; negative functional assays exclude HIT. (See 'Diagnosis' above.)

●Special populations – Spontaneous HIT and HIT in children are rare; evaluation is similar to HIT in adults. Hemodialysis and cardiopulmonary bypass often cause heparin-induced antibodies of unclear significance; we only test these individuals if suspicion for HIT is high. Evaluation of vaccine-induced immune thrombotic thrombocytopenia (VITT) is discussed separately. (See 'Diagnosis in special populations' above and "Management of heparin-induced thrombocytopenia (HIT) during cardiac or vascular surgery" and "COVID-19: Vaccine-induced immune thrombotic thrombocytopenia (VITT)".)

●Management – Management of HIT and VITT are presented separately. (See "Management of heparin-induced thrombocytopenia" and "COVID-19: Vaccine-induced immune thrombotic thrombocytopenia (VITT)", section on 'Management'.)

If HIT is excluded, we discontinue non-heparin anticoagulation, restart heparin if indicated, and evaluate other causes of thrombocytopenia. (See 'Differential diagnosis' above and "Diagnostic approach to thrombocytopenia in adults".)

ACKNOWLEDGMENT — The UpToDate editorial staff acknowledges Steven Coutre, MD (deceased), who contributed to an earlier version of this topic review.