Return to American College of Chest Physicians                                     Print This
Heparin-Induced Thrombocytopenia:
Is it an Underdiagnosed Condition?

An Overview of Heparin-Induced Thrombocytopenia: Frequency, Presentation, Diagnostic Challenges

Heparin-induced thrombocytopenia (HIT) is a complication in which heparin exposure leads to antibody formation and platelet activation. Devastating thrombosis can result, progressing to amputation or death. Yet HIT often goes unrecognized and may not be considered as a cause of thrombosis or thrombocytopenia, said Diane E. Wallis, MD, FCCP, of Midwest Heart Specialists in Downers Grove, IL. HIT can produce horrifying consequences. “The patient that got me interested...lost both arms and both legs,” she said. She described a young woman who lost a hand to HIT after receiving heparin flushes during dilation and curettage. “HIT is the most important and catastrophic complication of drug therapy today,” Dr. Wallis stressed.

Given the high odds ratio of thrombosis with HIT (OR: 36.9) compared to other thrombotic conditions (eg, OR: 24.4 for antithrombin deficiency or OR: 6.6 for Factor V Leiden) “it (HIT) should be number one in the differential diagnosis of anybody coming in with a thrombosis, particularly a post-admission or post-operative thrombosis,” Dr. Wallis emphasized. Yet, often when a patient with thrombosis is evaluated, “Nobody has looked to see if the patient was readmitted within 100 days, nobody has trended the platelet count, and nobody has included HIT in the differential diagnosis,” Dr. Wallis said.

One reason why physicians may not recognize HIT is that heparin use “exploded” in the US only since 1990 or so. At the Hospital of the University of Pennsylvania (HUP), heparin was used in 39% of the 36,000 admissions during 1 year, Dr. Wallis said (Personal communication, Gene Gibson, PharmD, associate director, Pharmacoeconomics, HUP). These figures exclude exposure through heparin flushes, which was not documented. “Patients hospitalized within the past year should be assumed to be heparin-exposed until proven otherwise,” she said.

Another problem may be the apparently contradictory data about the frequency of HIT, she said. One publication reported a rate of .39% patients with a positive serotonin release assay for heparin among 259 consecutive medical intensive care unit (ICU) and cardiac care unit patients (Verma AK, et al. Pharmacotherapy 2003;23:754). A review of several trials yields a rate of roughly 2% to 4%.

This discrepancy exists partly because the studies reporting higher rates were focused on certain populations. According to two published series, most cases of HIT occur among surgical patients (66% or 88%) (Warkentin TE, Kelton J. Am J Med 1996;101:502; Wallis DE, et al. Am J Med 1999;106: 629). Surgical patients with HIT were concentrated among those undergoing orthopedic (48%) and open heart (23% or 59%) procedures. Medical patients with HIT generally had “some devastating inflammatory process” such as cancer or systemic lupus erythematosus, Dr. Wallis said. Physicians should have a higher index of suspicion for HIT in such high-risk populations, she said.

The virtual absence of education about HIT also is an obstacle to its recognition. HIT receives perhaps one paragraph in major medical texts, while several pages are devoted to disorders that most physicians will never see. “It’s often not in the medical student curriculum and residents don’t hear about it on rounds,” she said.

The atypical presentation of HIT also confounds diagnosis. Traditional teaching holds that HIT presents after 5 days of heparin exposure. However, earlier presentation (as soon as 10.5 hours after heparin) occurs, especially in patients re-exposed to heparin within 100 days. One series reported this early-onset form of HIT in 30% of 243 patients (Warkentin TE, Kelton JG. N Engl J Med. 2001;344:1286).

Delayed-onset HIT also has been described. In two series, patients were discharged, then presented with HIT (mean 9.2 days after stopping heparin; Warkentin TE, Kelton JG. Ann Intern Med. 2001;135:502) or were readmitted with HITTS (heparin-induced thrombocytopenia with thrombotic syndrome) (mean of 14 days; Rine L, et al. Ann Intern Med. 2002:136:210). Speaking of HIT’s variable presentation, Dr. Wallis said, “I think the exceptions are more the rule than the exception.”

Other problems with the definition of HIT also confound diagnosis:
• Despite its name, HIT can occur without thrombocytopenia (Gerdisch M, Wallis DE, et al. Ann Hematol 2003; 82[Suppl 1]:S27; Warkentin TE. Semin Hematol 1998;35(4):9–16).
• Identifying another cause of thrombocytopenia does not preclude the presence of HIT.
• Some definitions of HIT require thrombosis, which can be difficult to detect.
• There are problems with the use of antibodies as a diagnostic criterion. For example, some patients are initially antibody-negative, and then become antibody-positive 5 to 14 days later. “There may be unexplained antibody…that we don’t detect,” Dr. Wallis explained.

Dr. Wallis suggested that the following clues should raise suspicion of HIT:
• Heparin re-exposure
• Sepsis, surgery, trauma
• Unusual thrombosis
• Platelet count decline of more than 50%, regardless of absolute count

Early detection of HIT and immediate cessation of heparin do not by themselves significantly reduce risk of thrombosis or death (Wallis DE, et al. Am J Med 1999;106:629-635). About 50% of patients develop thrombosis after heparin is stopped (Warkentin TE, Kelton JG. Am J Med 1996;101). “The only way I know to prevent HIT is not to use heparin at all, ever,” she said.

In the pathophysiology and treatment of HIT, Dr. Wallis said, “Think about HIT as a hematologic emergency, as a thrombotic storm. It’s the thrombin that has to be attacked.” Three direct thrombin inhibitors (DTIs) are available (bivalirudin, lepirudin, and argatroban). The latter two are approved for use in HIT. Data support the effectiveness of lepiruduin (Greinacher A, et al. Blood 2000;96:846–851) and argatroban (Lewis BE, et al. Arch Intern Med 2003; 163:1849) in HIT.

About 40% of patients develop antibodies to lepirudin (Greinacher A, et al. Circulation 2003;108:2062-2065). These antibodies were initially viewed as benign. However, there are now reports of anaphylaxis and death with re-exposure to lepirudin. Additionally, antibodies to lepirudin may increase the drug’s anticoagulant effect. Therefore, strict monitoring of activated partial thromboplastin time is necessary, Dr. Wallis noted.

Besides considering use of a DTI as alternative anticoagulation, HIT therapy involves removing all heparin exposure. “You have to be very vigilant,” she said. “I’ve had HIT patients go to the OR and the surgeon uses heparin-coated guide wires, catheters, and bathed prosthetic devices in heparin. You need to communicate with all the doctors involved in the care of this patient” to prevent heparin exposure. Dr. Wallis also suggested monitoring platelet counts until they are fully recovered. She counseled against prophylactic platelet transfusions, as they “add fuel to the fire.”

It is crucial to remember that warfarin is contraindicated in acute HIT unless a DTI is also given. Serious complications can result, including deep
vein thrombosis and venous gangrene (Warkentin TE, et al. Ann Intern Med 1997;127:804-812).

An Overview of Heparin-Induced Thrombocytopenia: Identifying the Most At-Risk Patients

John G. Kelton, MD, Dean, McMaster University Medical Centre, Hamilton, ON, Canada, defined HIT as a fall in platelet count to less than 100,000 to 150,000 within 5 or more days after starting heparin, with or without thrombotic complications, and with the exclusion of other causes of thrombo- cytopenia, plus a positive serologic test for HIT.

There are exceptions and caveats to this definition. In addition to those mentioned by Dr. Wallis, another is that some patients show a smaller drop in platelet count, especially in the presence of skin necrosis. Serologic testing is extremely sensitive but not specific. Therefore, “almost all HIT patients will have a positive test. However, many will have a positive test who do not have HIT,” said Dr. Kelton. No test specific for HIT is available now.

HIT typically presents with moderate thrombocytopenia (40,000 to 60,000), with or without thrombotic complications, Dr. Kelton said. Thrombotic complications may develop after presentation. Venous thrombi occur four times as often as arterial thrombi. There are “virtually no other illnesses that cause both venous and arterial thrombi, especially in the same patient,” Dr. Kelton said. Bleeding is rare. “My rule in thrombosis is always risk a bleed to prevent a thrombus. I’m always right,” Dr. Kelton said.

Risk of HIT rises with heparin dose, bovine (vs porcine) heparin, and unfractionated (vs low molecular weight) heparin (UFH vs LMWH). However, he cautioned, “HIT has been described for every heparin preparation given at any dose.” Orthopedic and general surgery patients receiving full-dose UFH are at higher risk for HIT than are those receiving heparin in coronary artery bypass graft surgery.

Despite this information about risk, Dr. Kelton said, “There is not enough data to say how often we should monitor our platelet count.”

“The ideal drug is thrombin inhibition,” he said. Both of the agents licensed in the US and Canada, argatroban and lepirudin, are highly effective and thrombin-specific (Figure 1). In addition to the concerns about lepirudin antibodies that Dr. Wallis mentioned, Dr. Kelton noted that lepirudin carries a risk of major bleeding roughly double that of argatroban.

Treatment of HIT involves stopping heparin, substituting a DTI, not using warfarin initially, then slowly introducing warfarin at low doses when platelet count rises. “Anticipate a 1-week transition” to warfarin, Dr. Kelton said. Appropriate duration of warfarin therapy is unknown, he said. He recommended a minimum of 3 months. Using warfarin without a DTI early in HIT can cause warfarin-induced necrosis due to the fall in protein C.

Thrombin inhibition should be used in persons with isolated thrombocytopenia as well as in those with thrombotic events, he said. This is because risk of progressive thrombosis in isolated thrombocytopenia is about 50%, according to prospective and retrospective studies, Dr. Kelton said.

Four prospective trials have been conducted using DTIs in HIT. All use historical controls. An analysis of those 4 trials (Hirsh, Heddle, Kelton. Arch Int Med in press) indicates that both thrombin inhibitors are highly effective at preventing new thromboembolic events (TEs) in patients with isolated HIT. Neither agent significantly reduces rates of death and amputation. “The death rate in HIT is still unacceptably high,” Dr. Kelton declared, citing rates of 10% to 20%, depending upon whether deaths of patients both directly and indirectly attributable to HIT are included (Hirsh, Heddle, Kelton. Arch Int Med in press).

In an effort to determine risk factors for poor response to DTIs, Dr. Kelton and colleagues applied 14 predictors of negative outcome from the literature to data from two prospective trials of argatroban (Lewis BE, et al. Circulation 2001:103:1838-1843; Lewis BE, et al. Arch Intern Med 2003:163:1849-1856). Analysis of the first trial revealed that 7 of the 14 variables were associated with poor outcome. Attempting to validate these findings in the second trial demonstrated that platelet count predicted risk of the composite endpoint (thrombosis, amputation, and death) (P < .001). Renal impairment predicted risk of death (P < .007). Cardiovascular surgery predicted risk of amputation (P < .001). “The weakness of this model is that it is using correlation coefficients to predict causality,” Dr. Kelton noted. Using the second trial to validate the findings of the first trial is an attempt to compensate in part for this weakness.

Based on these findings, Dr. Kelton said, “The lower the platelet count, the more you should be thinking about additional therapy besides the DTI. This could include a DTI plus an antiplate-let agent plus high-dose intravenous immunoglobulin G (IgG).

Using LMWH as much as possible also can reduce risk of HIT. “The diagnosis depends on a high clinical awareness,” Dr. Kelton stated.


Heparin-induced Thrombocytopenia in the Cardiopulmonary Patient

Victor F. Tapson, MD, FCCP, of Duke University Medical Center, Durham, NC, opened his talk with a case of HIT in a 65-year-old male with chronic obstructive pulmonary disease and heart failure who presented with fever, worsening cough, dyspnea. He died on day 12 of hospitalization. HIT serology was evaluated the day before he died. His platelet count on day 4 was half that of admission, and he had heparin exposure 3 weeks earlier. “I’m not sure how many of us would pay attention to that early platelet count drop to 146,000 (on day 4),” he said.

The patient received heparin after this point, and died with thrombosis throughout his body — “massive” pulmonary embolism, renal vein thrombosis, aortic thrombosis, femoral/ popliteal arterial thrombosis, and others. This is a typical presentation of HIT “except for extreme thrombosis in arteries and veins,” Dr. Tapson said. Given how commonly heparin is used and how immunogenic it is, “You’d think we’d recognize this (HIT) very easily,” he said.

Heparin is one of several reasons for thrombocytopenia in the ICU, and one of several drugs that can cause it. A retrospective study of critically ill patients found that heparin exposure and/or hemodynamic instability appeared to be the strongest correlates with thrombocytopenia (Bonfiglio MF, et al. Ann Pharmacother 1995;29:835-842).

HIT in Cardiopulmonary Bypass
Analysis of three prospective and four retrospective trials reveals that HIT develops in about 2% to 2.4% of cardiopulmonary bypass (CPB) patients. “Unfractionated heparin during cardiopulmonary bypass is remarkably immunogenic,” Dr. Tapson said. About 25% to 50% of CPB patients develop heparin-dependent antibodies over 5 to 10 days. From 7% to 40% of these seroconversions include high-titer IgG antibodies that activate platelets in vitro (Pouplard C, et al. Blood 2002;100:16a-17a; Warkentin TE, et al. Blood 2000; 96:1703-1708; Bauer TL, et al. Circulation 1997;95:1242-1246). Platelet activation can lead to coagulation and HIT. A minority (less than 10%) of antibody-positive patients develop thrombocytopenia, even when heparin is continued through the postoperative period. The risk of HIT is 1% to 3% if UFH is continued throughout the postoperative period. In general, patients with the strongest antibody test results are most likely to develop HIT (Warkentin T.E., Heddle N.M. Curr Hematol Rep 2003; 2:148-157).

Because of the high incidence of antibody formation and the potential for late-onset HIT, patients readmitted after coronary artery bypass graft surgery with symptoms of thrombosis should not receive heparin until results of antibody testing become available, he said.

Thrombosis is common in HIT (40% to 75%). In the HIT population overall, venous thrombosis is more common than arterial thrombosis. Postcardiac surgery patients with HIT experience arterial thrombosis more frequently than do other HIT patients (Warkentin TE, Greinacher A. Ann Thorac Surg. 2003;76:638-648). Placement of mechanical devices such as catheters and balloons in the vasculature may promote thrombosis, he said. “If someone’s had a balloon stuck in his femoral artery a couple of times,” arterial thrombosis is “not surprising,” he said. However, “it’s worth taking note of the platelet count, realizing it could be a case of HIT thrombosis,” Dr. Tapson advised.

HIT may be “fairly common” after placement of an intra-aortic balloon pump (IABP), said Dr. Tapson. One publication reported the incidence of HIT among patients with IABP at 4.5% (Walls JT, et al. ASAIO J 1992;38: M574-6). Nearly half of those with HIT experienced thromboembolic complications (48.6%), and nearly half died in the hospital (42%).

Dr. Tapson said that his colleagues prefer to use heparin for anticoagulant therapy during cardiopulmonary bypass. Therefore, delaying surgery for more than 100 days after an episode of HIT “might be ideal”. That way, heparin can be used, as the heparin-dependent antibodies should no longer be present in the body. If surgery cannot be delayed, a protocol for using lepirudin during CPB is available (Warkentin TE, Greinacher A. Ann Thorac Surg 2003;76:638-648).

HIT was diagnosed in 1.9% (82/4261) of open heart surgical patients evaluated from 1981 to 1991 (Walls JT, et al. Ann Thorac Surg 1992;53: 787–791). If this rate sounds low, “think of all the cardiac surgery cases done in the US, Canada, Europe every day,” Dr. Tapson said. Those in whom heparin-dependent antibody was identified preoperatively (n = 12) did not receive postoperative heparin. They experienced no TE events and no deaths. Patients in whom heparin-dependent antibody was detected postoperatively (n = 70) received heparin for varying periods after surgery. About 44% experienced TE complications, leading to death in 33% (23/70).

The formation of a registry for HIT, called Complications After Thrombo-cytopenia Caused by Heparin Registry (CATCH), should yield more data about the incidence of HIT and thrombosis in patients treated with UFH or LMWH.

Dr. Tapson recommended considering early DTI therapy. “We don’t necessarily have to wait until we have catastrophic thrombosis to start therapy,” he said. Precisely when to begin treatment “is better left to bedside judgment,” he said. The CATCH data should produce more information about when to begin DTI in HIT. He also urged physicians to consider HIT in diagnosis. “Anytime someone has some degree of thrombocytopenia and some degree of thrombosis anywhere, we have to think about HIT,” Dr. Tapson said.


Return to American College of Chest Physicians                                    Print This

All contents Copyright © 1999 - 2004 Medical Association Communications