Heparin can lower the platelet count through two mechanisms: a direct effect on platelets that occurs within 2 days, usually resolves despite continued therapy, and is of no clinical consequence, or immune heparin-induced thrombocytopenia (HIT).

The information below covers only (immune) HIT, a life-threatening condition with a high thrombosis risk caused by platelet-activating HIT antibodies. HIT usually develops between days 5-10 days of heparin therapy, but it can occur immediately if preformed antibodies are present from a recent heparin exposure.

HIT is an immune reaction to platelet factor 4 (PF4) complexed with heparin, which results in an intensely prothrombotic state. HIT is best viewed as a clinicopathologic syndrome (because neither thrombosis/thrombocytopenia, or HIT antibodies, are sufficient for diagnosis in isolation), and as a spectrum of disease.

Continue Reading

Many patients treated with heparin develop anti-PF4/heparin antibodies (HIT antibodies), but only some, particularly those with higher titers of antibodies, develop complications. In these patients, HIT antibodies bind PF4/heparin complexes which have bound to activated platelets, stimulating the platelets to release more PF4, resulting in a positive feedback loop.

Affected platelets are prone to aggregation, which results in their removal from circulation and thrombocytopenia. In many patients, the intense platelet activation results in thrombosis. Thromboses may be typical venous, fibrin-rich clots, or arterial, platelet-rich “white” clots.

While HIT is uncommon, recognition is critical so that the inciting heparin exposure can be stopped and alternative anticoagulation substituted to treat or prevent thrombosis. Even with prompt recognition and treatment, HIT patients may suffer morbid or fatal thromboembolism or suffer other complications such as skin necrosis, stroke, or limb loss.

Clinically, HIT is more common in surgical patients (compared to medical patients), women, and in those receiving unfractionated heparin (UFH) compared with those receiving low molecular weight heparin (LMWH).

A clinical risk scoring system (the 4Ts) helps predict the risk of HIT in various clinical settings. In addition, two types of tests are available for HIT: ELISA assays for the HIT antibodies, and functional assays which demonstrate the activity of these antibodies on platelets.

You cannot rely on laboratory assays or clinical events alone to make the diagnosis. In general, a positive functional assay should be required for diagnostic confirmation, although high clinical suspicion plus a positive HIT antibody assay may be sufficient in some cases (see “Default Management” for a discussion of treatment needed during the diagnostic process).

The first step is to consider the pretest risk of HIT by using a validated risk scoring system. The most widely used is the 4Ts score:

(T)hrombocytopenia: if the platelet count:

  • falls above 50% with a nadir greater than 20,000, assign 2 points

  • falls 30-50% or nadirs between 10 and 19,000, assign 1 point

  • falls below 30 percent or nadirs less than 10,000, assign 0 points

(T)iming: if the onset occurs:

  • between days 5-10 of heparin therapy, or within 1 day if the patient received heparin in the last 30 days, assign 2 points

  • possibly between days 5-10 but not clear due to missing data points, or begins after day 10, or begins within a day if the last heparin exposure was 30-100 days, assign 1 point

  • within 4 days without recent exposure to heparins, assign 0 points

(T)hrombosis and other sequelae: if the patient suffers:

  • new thrombosis, skin necrosis, or acute systemic reactions after intravenous heparins, assign 2 points

  • worsening or recurrent thrombosis, erythematous (nonnecrotic) skin lesions, or unproven thrombosis, assign 1 point

  • no effects, assign 0 points

o(T)her causes of thrombocytopenia

  • if none are present, assign 2 points

  • if possible causes are present, assign 1 point

  • if definite other causes are present, assign 0 points

Probability of HIT by total score

  • 0-3: low (<1-4% risk)

  • 4-5: intermediate (8-29% risk)

  • 6-8: high (21-100% risk)

Patients with a low 4T score have a HIT antibody prevalence of less than 1 to 4%, and are highly unlikely to have positive functional assays (e.g., none of 458 patients in one series); thus, such patients can generally continue with heparin therapy and monitoring.

Patients with intermediate and high scores generally require testing, which means that their heparin therapy must be interrupted (tests involve the addition of heparin during the test, so samples must be heparin-free when provided), and they may require alternate anticoagulation for a pre-existing condition or because of a high suspicion of HIT while awaiting test results.

Other scoring systems include the HEP (HIT Expert Probability) Score and the Lillo-Le Louët model. The HEP score has been found to perform similarly to the 4T score, but is limited by lower familiarity and increased complexity.

Regardless of the scoring system used, confirmatory laboratory tests are required to diagnose HIT. Three tests are commonly used: two functional tests (the gold standard serotonin release assay and heparin-induced platelet aggregation [HIPA]), and the ELISA for HIT antibodies. The performance of all three tests has varied depending on the clinical situations and/or the experience of the laboratory.

1) In the serotonin release assay, normal platelets are radio-labeled with 14C-serotonin and exposed to patient serum and varying amounts of heparin. While this test is highly accurate, with a sensitivity and specificity of over 95% (ranges 88-100%), it is more expensive, requires radioactive reagents, and is not available at many centers.

2) HIPA assays entail the addition of platelets from a normal donor to serum or plasma from the patient, followed by the measurement of platelet aggregation in the absence of heparin and with low and high heparin concentrations. A positive test is highly suggestive of HIT with a specificity of about 91% (range 77-100%). The sensitivity of HIPA, however, is operator dependent and may be as low as 39-81%, although it has been reported to be over 90% at experienced laboratories.

3) The ELISA assay for HIT antibodies is useful for excluding HIT. Patient serum is added to heparin/PF4 complexes, and if HIT antibodies are present and bind, they are detected by a second antibody. ELISA offers high sensitivity (>97%) but limited specificity (74-86%). The negative predictive value of the test is over 95%.

HIT may present in a classic or atypical fashion. In its most obvious presentation, a case of HIT would involve a moderate, new onset thrombocytopenia (either >50% decline, or to <150,000, with a nadir above 20,000) without other known etiology, that develops after 5-10 days of exposure to heparins, or promptly, if a patient has received heparin in the last 3 months or so and still has pre-existing antibodies (for this reason, a platelet count within 24 hours is advised after heparin re-exposures).

For the purposes of HIT suspicion, declines in platelet count should be compared to either the pre-heparin count, or the highest count since heparin initiation that precedes the suspicion-generating fall in counts. HIT patients may also develop, or present with, arterial or venous thrombosis. Thromboses are most often lower extremity deep venous thrombosis (DVT) or pulmonary embolism (PE) but may involve any vascular bed, especially at sites of injury, and findings may include upper extremity DVT, arterial thrombosis with limb ischemia, or unusual visceral or cerebral thromboses.

A patient with pre-existing antibodies to heparin can develop an acute reaction on re-exposure to heparins, including anaphylactoid reactions, including fever, chills, tachycardia, dyspnea, or even cardiopulmonary arrest.

However, HIT can also present atypically. The development of any of the following may also represent HIT:

  • HIT complicating other expected thrombocytopenia, e.g. patients with sepsis or post-cardiopulmonary bypass. Follow the pattern of thrombocytopenia for clues, such as a drop and recovery after surgery with a second (unexpected) fall.

  • Necrosis or other heparin injection site reactions without thrombocytopenia.

  • Less than 50% decline in platelet count.

  • Thrombosis developing on heparin therapy, before thrombocytopenia (33% of cases in one series; another 26% had simultaneous thrombosis and thrombocytopenia).

  • Thrombocytopenia after longer term heparin receipt, as with a dialysis patient.

  • New thrombosis or thrombocytopenia AFTER the cessation of heparin, in occasional patients who develop heparin-independent platelet activating antibodies. In one series of patients, this occurred on an average of 9 days (range, 5-19) after heparin cessation.

HIT may be underappreciated and its recognition delayed. In one series, physicians failed to stop heparin or test thrombocytopenic heparin recipients, more often than not. However, thrombocytopenia developing after 1-2 days of heparin therapy is NOT immune HIT, provided the patient has not been recently exposed to heparins.

HIT occurs more often in surgical patients than in medical patients (relative risk (RR) 3.2), and the highest risk occurs amongst orthopedic patients, particularly those on prolonged courses of UFH. HIT is uncommon in dialysis recipients.

While UFH poses an increased risk of HIT compared to LMWH (estimates vary from a RR of 5.3 to a 10-fold increase), LMWH can certainly induce complications of HIT in patients in whom HIT antibodies formed after UFH exposure. HIT appears to be rare under the age of 40.

Reported rates of ELISA positivity for HIT antibodies (and reported rates of HIT) vary by the population and type of exposure:

  • orthopedic surgery patients on UFH: 14% (2.6-5%)

  • cardiac surgery patients on UFH: 15-70% (1-2%)

  • general medical, neurology, cardiac catheterization, or acute dialysis patients on UFH: 8-20% (0.8-3%)

  • general pediatric, obstetrics, or hemodialysis patients on UFH: 0-2.3% (0-0.1%)

  • general medical, surgery, neurology or orthopedics patients on LMWH: 2-8% (0-0.9%)

  • Pregnant women and general pediatrics patients on LMWH: unknown (0-0.1%)

Other reports suggest a rate of 0.2-5% of patients treated with heparins for over 4 days, with an overall rate of 2.6%. Prospective observational studies have reported rates of 50% platelet count declines of 15-42%. These published reports suggest a risk of HIT that exceeds the rate expected based on the clinical experience of many hospitalists and known activity of heparins. In other words:

Practicing hospitalists may wish to estimate the risk of HIT associated with heparin exposure by considering both published rates of HIT and local rates, which hematology consultants may know, as variable reported rates make it difficult to know how worried to be about the rare complication of HIT when dealing with the common issues of thrombocytopenia and heparin receipt.

New onset thrombocytopenia can occur with many causes such as sepsis, thrombotic thrombocytopenic purpura / hemolytic uremic syndrome, disseminated intravascular coagulation, after bypass surgery or aortic balloon pump therapy, a variety of other medications, or bone marrow disorders or splenic sequestration; critical illness of many types is a potent risk factor for both thrombocytopenia and VTE. Platelet clumping may mimic thrombocytopenia.

New or worsening thrombotic events despite prophylaxis or therapeutic anticoagulation may occur due to immobilization/paralysis, surgery, cancer, central lines, multiple traumas, general medical conditions, or the interplay of these and other risk factors. Hospitalists should definitely worry about HIT when a patient on therapeutic heparins has a new VTE event and any time there is an atypical thrombosis (e.g., arterial) on heparins.

Patients with anti-thrombin III (AT-3) deficiency are particularly prone to heparin failure (as heparin requires AT-3 for activity). Obese patients may receive less protection from standard heparin doses, and UFH protocols are difficult to follow to the letter, resulting in underdosing of some patients. Anti-phospholipid antibody syndrome and Trousseau’s syndrome can cause particularly prothrombotic states.

Heparin infusion reactions could be mistaken for infection, with fever, rigors and tachycardia, but if there is any chance heparin is the cause, the infusion should be stopped immediately and platelet count checked. Injection site reactions might be confused with bruising, and skin necrosis could be mistakenly attributed to warfarin or emboli from the disorder under treatment, e.g. atrial fibrillation.

HIT typically causes either no abnormalities in the physical exam, or if complicated by VTE, would cause findings consistent with VTE from other causes (limb pain, swelling, and redness from DVT, or tachycardia, tachypnea, hypoxemia, and hypotension from PE).

Arterial thrombosis may be manifest by signs of ischemic stroke, or critical limb ischemia with diminished pulses, pallor, paralysis and paresthesia, and coolness. Necrosis at heparin injection sites is highly suggestive. Bleeding and petechiae are unusual because the platelet count is seldom low enough to compromise platelet function.

See above in section II.

See above in section II.

The 9th American College of Chest Physicians (ACCP) guidelines make a 1C recommendation for ultrasound of the lower extremity veins, with or without signs or symptoms of thrombosis, in the setting of newly diagnosed HIT.

Because the ELISA assays for HIT antibodies have a low specificity and positive predictive value, do not screen heparin recipients for HIT apart from monitoring platelet counts and monitoring for signs and symptoms of heparin reactions or new thromboses. The 9th ACCP guidelines suggest limiting platelet count monitoring to patient populations with a greater than 1% risk of HIT.

Similarly, estimate a patient’s pre-test probability of HIT before initiating any laboratory work-up by using the 4Ts as described above, as indiscriminate testing will lead to interruptions in heparin therapy, increased cost from testing and heparin alternatives, possible bleeding risks from less familiar and predictable anticoagulants, and false positive diagnoses. A patient who is low probability for HIT on the basis of their 4T score (0-3 points) has a less than 1-4% chance of having significant levels of HIT antibodies, and a close to zero chance of having a positive functional assay or complications of HIT. One study looked at incorporating the 4Ts into a physician order entry system that would require the physician to review the 4Ts prior to ordering HIT laboratory work. Reviewing the 4Ts resulted in less testing for HIT.

A patient with a low probability of HIT before or after testing should have their anticoagulation regimen reconsidered; treating them as if HIT is present may be appropriate if their risk is felt to be sufficiently high. Alternatively, if heparins are continued, vigilance for platelet count drops and thrombosis may be appropriate.

A patient in whom HIT has been confirmed or in whom there is a higher likelihood of HIT requires two key interventions: the immediate cessation of all exposure to heparins, and alternative anticoagulation. Heparin exposure (UFH, LMWH, flushes, or continued use of any heparin coated lines or devices) must be stopped to avoid continued stimulation of HIT antibodies.

Heparin should be listed as an allergy, and signs should be posted at the bedside to avoid inadvertent use of heparin flushes during routine care. An alternate anticoagulant must be provided because HIT patients have a very high risk of thrombosis during the acute phase: roughly 50% at 30 days, with most of the risk concentrated in the first 10 days and nearly all of it in the first 21. Observational studies show some physicians omit this alternate anticoagulation, perhaps because it can be inconvenient, costly, or because there is no thrombosis (yet). However, most clinicians follow guidelines to treat unprovoked DVT >3 months, and the risk of skipping this therapy is less than a fifth of the VTE risk with acute HIT.

Unless contraindicated, anticoagulation must continue until the prothrombotic state is resolved, or as appropriate for any thrombotic complications that develop. More research is required to determine the optimal duration of anticoagulation in the absence of thrombosis, but based on the timing of thromboses, therapy for 3-4 weeks should be considered.

Additional immediate management includes the use of vitamin K if HIT is diagnosed during warfarin therapy. The 9th ACCP guidelines recommend (Grade 2C) vitamin K (consider 10 mg of oral [or 5-10 mg IV] vitamin K) in this situation, to allow production of anticoagulant factors; simultaneous alternate anticoagulation is important.

Argatroban is the primary agent used for the initial treatment of HIT (1C recommendation from the ACCP guidelines), as the other two suggested drugs (lepirudin and danaparoid) are not available. Argatroban is a direct thrombin inhibitors (DTI), in contrast to heparin, which inhibits thrombin via activated antithrombin III. Because most physicians use these agents infrequently, and the titration is complex, use of institutional protocols and pharmacy and/or hematology consultation are advised.

There are no randomized, controlled trials or head to head trials to compare these agents, which were tested against historical controls. Thus, the decision on which agent to use depends on availability and the presence or absence or hepatic or renal failure. More data is needed.

The initial dose for patients without liver impairment is 2 mcg/kg/min. A steady state is achieved in 1-3 hours, and the effect can be estimated by monitoring the aPTT. The goal prolongation in the aPTT is 1.5-3 times the baseline value, not to exceed 100 seconds. Adjustment of the argatroban dose up to 10 mcg/kg/min is used to achieve therapeutic aPTT values.

No dose modification is required for patients with renal failure. Patients with impaired hepatic function or who have heart failure, anasarca, multiple organ failure, or are post-cardiac surgery may be treated with argatroban starting at 0.5 to 1.2 mcg/kg/min, although consideration should be given to the substitution of an alternate agent in this population. Previously, the preferred agent was lepirudin; now, bivalirudin is preferred in patients with liver dysfunction.

Argatroban therapy costs about $1,000 per day, and at least 5 days of therapy AFTER platelet counts have recovered should be anticipated (for overlap with coumadin; see below in “Long-term management”).

Lepirudin is a renally-cleared, recombinant version of the leech-derived DTI hirudin. Formerly an option for HIT patients with impaired hepatic function, lepirudin is no longer available. Bivalirudin may be the most appropriate agent for patients with liver failure.

Bivalirudin is a hirudin analog most commonly used during percutaneous coronary intervention (PCI). It has been approved for PCI in patients with HIT, but not for the treatment of HIT itself. However, it has been used in HIT patients (starting dose 0.15 mg/kg/hr) as well as with either liver (starting dose of 0.14 mg/kg/hr has been used) or in renal failure with or without liver failure (starting doses 0.03-0.05 mg/kg/hr were effective). Doses are adjusted to achieve an aPTT 1.5-2.5 times the baseline value. In one small study, a pharmacy-driven protocol for bivalirudin was used to treat patients with HIT. Bivalirudin recipients reached a therapeutic aPTT faster and had less bleeding outcomes than patients being treated with argatroban. Further research is needed to clarify the role of bivalirudin in HIT.

Other agents:

Fondaparinux is a synthetic pentasaccharide which activates antithrombin 3, similar to heparins. While HIT antibodies have been detected after treatment with fondaparinux, HIT does not occur or is vanishingly rare. The drug has a predictable response and does not require monitoring, while intravenous DTIs require careful monitoring and adjustment. Thus, it is preferred to intravenous DTI by some experts. Fondaparinux received a 1C recommendation for HIT in the 2008 ACCP guidelines, but in the 2012 update, it was recommended only for: 1) pregnant patients when danaparoid is unavailable, and 2) patients with a history of HIT and acute thrombosis NOT related to acute HIT (2C recommendations). Increasing experience with this drug may lead to increased use in the future. The drug has a long half life, cannot be reversed, and is not Food and Drug Administration (FDA) approved for HIT.

Four new target-specific, non-heparin, oral anticoagulants are indicated for the treatment of VTE: the DTI dabigatran and the FXa inhibitors rivaroxaban, apixaban, and edoxaban. Some in-vitro data suggests that dabigatran and rivaroxaban do not interact with PF4 and may be treatment options for HIT (apixaban and edoxaban likely act similarly). These medications would be simpler, more convenient, and far cheaper than intravenous DTI, but clinical experience is limited. Other limitations include renal clearance and increased risk with impaired hepatic function. Dabigatran also required initial treatment with parenteral anticoagulants in VTE management trials. A multicenter study called “Design of the rivaroxaban for heparin-induced thrombocytopenia study” is currently underway.

The heparin-like compound, Danaparoid, is unavailable in the USA, but is used elsewhere, and received a 1C-2C recommendations in the 9th ACCP guidelines. It carries a 3.2% rate of cross reactivity with heparin. As with the DTIs, there is no direct comparison data to judge its relative safety or efficacy.

Immune globulin G (IVIG-G) blocks platelet Fcγ receptors, and limited data suggests that adding IVIG-G to anticoagulation might benefit patients at a high bleeding risk because of conditions like pregnancy, or autoimmune HIT. If used, the dose is 2 g/kg over 2 days.

Routine physical exams in HIT patients should search for edema or discoloration that suggests upper or lower extremity DVT, pulses for arterial thrombosis (although this is likely to be highly symptomatic in an awake patient), and signs of bleeding. Stay alert for less common manifestations such as the focal neurologic findings of stroke, or alteration in mentation and level of consciousness that may occur with cerebral sinus thrombosis.

Patients with HIT should receive a daily CBC with platelet count until it is stably improved, as well as laboratory monitoring of their anticoagulation, which would typically include aPTT testing as described under “immediate management.” Fondaparinux, and the new target specific oral agents, require no coagulation studies for monitoring.

The standard of care for long-term anticoagulation in HIT is warfarin. However, in coming years, other oral anticoagulants are likely to earn a role in HIT treatment.

Warfarin’s advantages include the most experience in HIT therapy, physician familiarity, the ability to treat patients with renal failure, and the possibility of reversal with plasma or vitamin K. However, there are many drug-drug and drug-diet interactions, unpredictable responses to therapy, and issues with warfarin sensitivity due to malnutrition, liver disease, debility, and other conditions making warfarin less attractive in some settings.

Most importantly, warfarin must not be started until platelet counts have recovered to at least 150,000 and ideally to baseline (Grade 1C recommendation) because warfarin rapidly depletes the anticoagulant factors protein C and S, and patients should not be exposed to this increased thrombotic risk until their HIT-related thrombosis risk is diminished.

Avoid high initial doses, using no more than 10 mg; using no more than 5 mg is probably best in most circumstances unless a prior effective dose is known, if only because this is the best studied dose in hospitalized patients. Similarly, most experts advise that DTI should be overlapped with warfarin for a minimum of 5 days with 2 consecutive therapeutic international normalized ratio (INR) values, just as heparin and warfarin must be overlapped in non-HIT patients. Thrombotic complications have been reported when less than 5 days of overlap have been used. The goal INR is 2-3.

To transition from argatroban, co-administer warfarin after platelet recovery, and check a daily INR (on doses of up to 2 mcg/kg/min; if using a higher dose, reduce to 2 mcg/kg/min for 4-6 hours before checking the INR). Once the INR is greater than 4 on combined therapy, discontinue argatroban and re-check the INR in 4-6 hours. If the INR is sub-therapeutic, resume argatroban and repeat the procedure in 24 hours. If the INR is therapeutic, the drug can be discontinued and warfarin alone administered (with the caveat that 2 consecutive daily INRs may reduce the risk of thrombosis in early warfarin therapy).

Fondaparinux, if chosen for the initial management of HIT, could be continued for the duration of anticoagulation, but conversion to an oral anticoagulant would be preferred for improved cost, comfort, and convenience.

Apixaban, rivaroxaban, and edoxaban are convenient, orally active anticoagulants that don’t require monitoring or cross titration with parenteral agents, but have not been studied for use in HIT. Similar considerations apply to dabigatran, but it appears to need an induction period with a parenteral agent. Converting patients to these drugs is easy because they work within an hour or two, they can simply be substituted for the parenteral agent when the next dose is due, and substituted for warfarin before the INR is subtherapeutic. Converting patients to other agents is more complicated and depends on renal function; see the package inserts or consult pharmacy for details. See other considerations in their use in “Immediate management”.

Warfarin must be reversed or withheld until the platelet counts recover because it depletes anticoagulant factors and increases the risk of thrombosis.

Avoid platelet transfusions which may increase the risk of thrombosis. If truly necessary (e.g., for bleeding or a planned invasive procedure), clinical experience suggests they are safe (Grade 2C recommendation).

Obtain pharmacy consultation and carefully administer intravenous DTI, which have complex dosing and monitoring parameters.

Options in renal insufficiency include argatroban, the best studied and approved agent, and bivalirudin (see dosing under “immediate management”).

Argatroban is cleared by the liver, and the starting dose must be adjusted in the setting of liver dysfunction. 0.5 up to no more than 1.2 mcg/kg/min is an appropriate starting dose in patients with meaningful elevations in their bilirubin or INR due to reduced synthetic function. However, patients with reduced liver function may be more safely treated with bivalirudin.

The lower starting argatroban dose of 0.5-1.2 mcg/kg/min is also advised for patients with either congestive heart failure or anasarca, or after cardiac surgery. In critically ill patients with multi-organ dysfunction, an even lower starting argatroban dose of 0.2 mcg/kg/min may be appropriate, although this is based on limited data.

No change in standard management.

No change in standard management.

No change in standard management


No change in standard management.

No change in standard management.

Expect increased warfarin effects in malnourished patients or those with reduced liver synthetic function. Expect increased argatroban effect in patients with reduced liver synthetic function, and consider using bivalirudin as an alternative.

See more specific headings throughout the chapter.

Anticipate involving family members in the anticoagulation management of cognitively impaired patients, as warfarin therapy requires attention to drug-drug and drug-diet interactions and conscientious follow-up.

Prominently note the patient’s allergy to heparin. Caution covering physicians about the potential risk of thrombosis and include a suggested management plan (e.g., “consult vascular surgery immediately if there is any suggestion of arterial thrombosis”). Include suggestions for the management of complex and rarely used anticoagulants, like argatroban, if applicable.

Length of stay is hard to estimate because it may be determined by the patient’s underlying reason for admission, and because recovery in platelet counts occurs over a broad time range (4-14 days). Most patients require hospitalization during this time, plus an additional 5 (and sometimes more) days for warfarin to be overlapped with intravenous DTI. Using subcutaneous (i.e., fondaparinux) instead of intravenous anticoagulants might permit earlier discharge in patients who are only awaiting a therapeutic INR on warfarin.

Prior to discharge, patients must have recovered their platelet counts, been successfully transitioned to an outpatient anticoagulant (most commonly warfarin, with a therapeutic INR; occasionally fondaparinux), received appropriate heparin allergy and anticoagulation teaching (in accordance with National Patient Safety Goal 3E), and received a follow-up plan which includes laboratory follow-up of their anticoagulation regimen and a proposed duration of therapy, which depends on whether or not thrombosis occurred.

Treatment should continue for at least one month without thrombosis, with some experts advising and variably suggesting 1-3 months of therapy, and at least 3-6 months in the presence of thrombosis. Here, 3 months represent a minimum only somewhat comparable to a provoked VTE, as the prothrombotic state of HIT cannot be said to be as transient as typical provocations like surgery or a removed central line.

See below.

HIT patients should typically be seen in a hematology clinic for follow-up, and usually require INR follow-up for their warfarin therapy. Patients may require an INR check within a few days if they were discharged as soon as they were therapeutic, or within 7-10 days if they were stably anticoagulated in the absence of significant changes to their medications and clinical status.

No specific studies are required beyond the establishment of platelet count recovery (occurs BEFORE warfarin can be started) and the achievement of therapeutic anticoagulation. Clinical assessment for bleeding, or hemoglobin monitoring if there is a concern bleeding has occurred, is appropriate for all inpatients on anticoagulants.

Most experts would obtain coagulation studies, as well as a follow-up CBC; basic chemistries may be useful if the patient is receiving a renally-cleared anticoagulant (fondaparinux, dabigatran) and changes to their renal function are possible.

Even uncomplicated acute HIT represents a period of high VTE risk as well as a condition which is usually treated with intravenous DTI, requiring inpatient monitoring and adjustment. The DTI must be continued until the platelet count has recovered, and only then can one begin warfarin treatment, which itself will require at least 5 days of therapy in combination with DTI, and possibly longer.

Several alternative therapies offer the possibility of reduced hospital lengths of stay: fondaparinux can be administered subcutaneously outside of the inpatient setting, during the overlap period with warfarin. There is experience with fondaparinux therapy for HIT, and it is recommended over intravenous therapies by some experts, but it is not FDA approved for this indication (nor is it approved for VTE therapy without transition to warfarin). Use of target specific oral anticoagulants are another option (see “Immediate management” for details).

A hospitalist contemplating either of these agents might be able to discharge a patient without HIT complications or other ongoing medical conditions substantially earlier, but would have to consider that clinical experience in HIT treatment generally entails the longer stays associated with IV DTI and conversion to warfarin, and this period of observation may be beneficial to patients given their high thrombosis risk.

Twenty to fifty-three percent of patients with HIT develop thrombosis, and a concomitant need for extended duration anticoagulation. Risk appears to be higher with platelet count drops of 70% or more, or higher titers of HIT antibodies. Depending on the site affected, patients may suffer limb loss (up to 5-10%), post-thrombotic syndrome, or sequelae of pulmonary emboli. Mortality rates between 8 and 35% have been reported.

After the cessation of heparin, platelet counts typically recover in 4-14 days, although some patients have a more prolonged recovery period and rare patients recover from (unrecognized) HIT and normalize their platelet counts despite ongoing heparin therapy. However, the thrombosis risk persists for days to weeks after platelet count recovery. HIT antibodies persist for a median of 85 days.

HIT patients should be carefully educated about their heparin allergy and the need to avoid even small doses of heparins in the future. While they may be able to receive cautious heparin therapy in the future, inadvertent heparin therapy while HIT antibodies are present would be particularly dangerous.

While the unique immune reaction of HIT may not actually recur with future exposure to heparin, there is often no need to risk re-exposure due to the availability of alternative anticoagulants, and any re-exposure should be carefully considered with the help of hematology consultation. Patients should receive education and warnings appropriate to their outpatient anticoagulation regimen, which is usually warfarin.

Joint Commission standards do not specify a particular approach to HIT management. However, elements of National Patient Safety Goal 3E (reducing harm from anticoagulant therapy) apply to the drugs used to treat HIT. This is particularly important for seldom-used and more difficult to titrate intravenous DTI, as well as warfarin therapy, for which appropriate pre-treatment testing, careful monitoring and protocolized adjustment, patient/family and staff education, and attention to drug-drug and drug-diet interactions are important.

Patients with possible HIT who require discontinuation of their UFH or LMWH prophylaxis should receive at least mechanical compression devices while their work-up is ongoing. Many experts would advise administration of an alternative drug for VTE prevention, but this decision rests on their pre-test probability of HIT.

If after a 4Ts assessment and consideration of your patient population and the risk associated with their heparin exposure, you determine that HIT is reasonably likely, full alternative anticoagulation should be given. Patients with a lower risk of having HIT could be treated with fondaparinux. Omission of pharmacologic prophylaxis should be reserved for patients who did not have compelling reasons for VTE prophylaxis in the first place, and who have a negligible risk of having HIT.

Prevention of HIT in medical patients can be accomplished in a variety of ways. When heparins are required, LMWH is safer than UFH, and more convenient. Fondaparinux can be substituted for heparins for many indications; fondaparinux may rarely induce HIT antibodies, but the risk of HIT with this agent is negligible. Also, the risk of HIT can be eliminated by the use of a target-specific oral anticoagulant instead of heparin. Four such drugs have been approved for treatment of VTE or prevention of stroke in nonvalvular atrial fibrillation: dabigatran, rivaroxaban, apixaban, and edoxaban (dabigatran and edoxaban require a course of heparin at the beginning of VTE therapy, while the others allow immediate oral therapy of VTE).

Warfarin and aspirin are used for VTE prophylaxis in orthopedic patient populations and pose zero risk of HIT, but have other limitations; in particular, aspirin may be accepted by some orthopedists and advocated by orthopedic professional societies, but the 9th ACCP guidelines consider it a second tier agent and recommend the use of LMWH in eligible patients. Target-specific anticoagulants may also be used for VTE prophylaxis after orthopedic surgery and pose no risk of HIT.

Prevention of complications of HIT can be accomplished by monitoring of platelet counts between days 4 and 14 of heparin therapy. The risk of HIT from LMWH is low enough to omit platelet counts for outpatients. Monitoring should occur within 24 hours of heparin therapy for patients who have received heparins in the last 3 months, because preformed antibodies may be present. Be sure to recognize relative thrombocytopenia (30-50% drops within the normal range) and consider HIT if a patient develops recurrent thrombosis despite anticoagulation with heparins.

The development of inpatient VTE despite prophylactic heparin is fairly common, however, because these agents are only about 50% effective for VTE prevention. Consider HIT and consider their 4T score, but treatment doses of heparins are still appropriate for most patients who fail heparin VTE prophylaxis.

Tardy-Poncet, B, Nguyen, P, Thiranos, JC. “Argatroban in the management of heparin-induced thrombocytopenia: a multicenter clinical trial”. Crit Care. vol. 19. 2015. pp. 396

Bain, J, Meyer, A.. “Comparison of bivalirudin to lepirudin and argatroban in patients with heparin-induced thrombocytopenia”. Am J Health Syst Pharm.. vol. 72. 2015. pp. S104-9.

Samuelson, BT, Glynn, E, Holmes, M. “Use of a computer-based provider order entry (CPOE) intervention to optimize laboratory testing in patients with suspected heparin-induced thrombocytopenia”. Thromb Res.. vol. 136. 2015. pp. 928-31.

Greinacher, A.. “Heparin Induced Thrombocytopenia”. N Eng J Med. vol. 373. 2015. pp. 252-61.

Warkentin, TE, Sheppard, JA.. “Serological investigation of patients with a previous history of heparin-induced thrombocytopenia who are reexposed to heparin”. Blood. vol. 123. 2014. pp. 2485-93.

Warkentin, TE, Sheppard, JA, Chu, FV, Kapoor, A. “Plasma exchange to remove HIT antibodies: dissociation between enzyme-immunoassay and platelet activation test reactivities”. Blood. vol. 125. 2015. pp. 195-8.

Krauel, K, Hackbarth, C, Fürll, B, Greinacher, A.. “Heparin-induced thrombocytopenia: in vitro studies on the interaction of dabigatran, rivaroxaban, and low-sulfated heparin, with platelet factor 4 and anti-PF4/heparin antibodies”. Blood. vol. 119. 2012. pp. 1248-55.

Joseph, L, Gomes, MP, Al Solaiman, F. “External Validation of the HIT Expert Probability (HEP) Score”. Thromb Haemost. vol. 113. 2015. pp. 633-40.

Kelton, JG, Arnold, DM, Bates, SM.. “Nonheparin Anticoagulants for Heparin-Induced Thrombocytopenia”. N Eng J Med. vol. 368. 2013. pp. 737-44.

Linkins, J, Dans, AL, Moores, LK. “Treatment and Prevention of Heparin Induced Thrombocytopenia: Antithrombotic Therapy and Prevention of Thrombosis”. Chest. vol. 141. 2012. pp. e495S-e530S.

Warkentin, TE.. “How I Diagnose and Manage HIT”. Hematology Am Soc Hematol Educ Program. vol. 2011. 2011. pp. 143-9.

Feinbloom, D, Srour, JF.. “Heparin Induced Thrombocytopenia”. Inpatient Anticoagulation. 2011.

Arepally, GM, Ortel, TL.. “Heparin Induce Thrombocytopenia”. New Engl J Med. vol. 355. 2006. pp. 809-17.

Warkentin, TE.. “Heparin Induced Thrombocytopenia: a ten-year retrospective”. Annu Rev Med. vol. 50. 1999. pp. 129-47.

Jang, IK, Hursting, MJ.. “When heparins promote thrombosis: review of heparin induced thrombocytopenia”. Circulation. vol. 111. 2005. pp. 2671-83.

Crespo, EM, Oliveira, GB, Honeycutt, EF. “Evaluation and management of thrombocytopenia and suspected heparin induced thrombocytopenia in hospitalized patients: The CATCH registry”. Am Heart J. vol. 157. 2009. pp. 651-7.

Lo, GK, Sigouin, CS, Warkentin, TE.. “What is the potential for overdiagnosis of heparin-induced thrombocytopenia?”. Am J Hematol. vol. 82. 2007. pp. 1037-43.

Lo, GK, Juhl, D, Warkentin, TE. “Evaluation of pretest clinical score (4 T's) for the diagnosis of heparin induced thrombocytopenia in two clinical settings”. J Thromb Haemost. vol. 4. 2006. pp. 759-65.

Greinacher, A, Farner, B, Kroll, H. “Clinical features of heparin induced thrombocytopenia including risk factors for thrombosis: a retrospective analysis of 408 patients”. Thromb Haemost. vol. 94. 1993. pp. 344-50.

Warkentin, TE, Sheppard, JI, Moore, JC. “Quantitative interpretation of optical density measurements using PF4-dependent enzyme-immunoassays”. J Thromb Haemost. vol. 6. 2008. pp. 1304-12.

Kiser, TH, Fish, DN.. “Evaluation of bivalirudin treatment for heparin induced thrombocytopenia in critically ill patients with heparin and/or renal dysfunction”. Pharmacotherapy. vol. 26. 2006. pp. 452-60.