A. Nonimmune Mechanisms in Heparin-Associated Thrombocytopenia
Klein and Bell (1974) reported on two patients who developed severe thrombocytopenia, thrombotic complications, and DIC, with hypofibrinogenemia and microangiopathic red cell abnormalities; i.e., these patients likely had severe HIT. This experience prompted Bell to perform the first prospective study investigating the frequency of thrombocytopenia complicating therapeutic-dose UFH (Bell et al., 1976). Sixteen of 52 patients (31%) developed a platelet count fall to less than 100 x 109/L, and some of these patients developed hypofibrinogenemia and elevated fibrin(ogen) degradation products. The authors speculated that a "thromboplastic contaminant" extracted along with heparin from beef lung could explain the thrombocytopenia. A subsequent randomized controlled trial by Bell and Royall
(1980) found the frequency of thrombocytopenia to be higher in patients who received bovine heparin (26%) compared with heparin of porcine intestinal origin (8%).
These investigators found no platelet-activating antibodies in plasma from the patients who developed thrombocytopenia (Alving et al., 1977), leading Bell (1988) to challenge the view that an immune pathogenesis explained HIT. However, as the Johns Hopkins group did not report thrombotic complications in any of their 37 patients who developed thrombocytopenia in their prospective studies, and given the apparent early onset of thrombocytopenia in many of their patients, it is likely that most of their patients did not have immune-mediated HIT.
B. Nonimmune (Type I) Versus Immune (Type II) HIT
A confusing situation arose. The terms "heparin-induced thrombocytopenia" or "heparin-associated thrombocytopenia" were often applied to any patient who developed thrombocytopenia during heparin therapy, whether presumed or proved to be caused by heparin-dependent antibodies or otherwise. Investigators in Australia, led by Chong (1981), also observed patients with thrombocytopenia in whom heparin-dependent, platelet-activating IgG antibodies could be identified. In a subsequent report that appeared in the Lancet, two distinct syndromes of "HIT" were described by Chong and colleagues (1982). The first, called "group 1," developed severe, delayed-onset thrombocytopenia with thrombotic complications in association with IgG antibodies that caused platelet activation. In contrast, "group 2" patients had mild asymptomatic thrombocytopenia of early onset.
In 1989, at a Platelet Immunobiology Workshop in Milwaukee, it was suggested to Chong that terminology describing these two types of HIT be formalized. Accordingly, Chong recommended the terms in a review article that appeared in Blut (Chong and Berndt, 1989), although (in reverse of the Lancet article nomenclature) the early, nonimmune disorder was named "HIT type I" and the later-onset, immune disorder referred to as "HIT type II." These terms subsequently became popular.
V. LABORATORY TESTING TO CHARACTERIZE THE HIT SYNDROME A. A Sensitive and Specific Platelet Activation Assay for HIT
Many clinical laboratories began to use platelet aggregation assays (Fratantoni et al., 1975; Babcock et al., 1976) to diagnose HIT. Problems with this type of assay, however, included low sensitivity (Kelton et al., 1984) as well as technical limitations in simultaneous evaluation of multiple patient and control samples. In 1983-1984, while working as a research fellow in the McMaster University laboratory of John Kelton, Dave Sheridan overcame problems of low test sensitivity by showing that washed platelets, resuspended in a buffer containing physiological concentrations of divalent cations, were very sensitive to platelet activation by HIT sera (Sheridan et al., 1986). The assay, known as the "platelet serotonin release assay (SRA)," was adapted from a method of platelet washing developed at McMaster University by the laboratory of Dr. Fraser Mustard. In particular, the emphasis on using physiological calcium concentrations was based on observations that "artifacts" of agonist-induced platelet activation were caused by the use of citrate anticoagulation resulting in low plasma calcium concentrations. One example of an artifact induced by citrate is that of two-phase aggregation triggered by adenosine diphosphate (ADP). At physiological calcium concentrations, only weak single-phase aggregation without thromboxane generation is triggered by ADP (Kinlough-Rathbone et al., 1983). Fortuitously, the washed platelet technique previously developed at McMaster University by Mustard and colleagues that Sheridan evaluated for its HIT serum-sparing properties rendered platelets far more sensitive to the platelet-activating properties of HIT antibodies than assays based on citrated platelet-rich plasma. Modified washed platelet assays have subsequently been developed by other investigators (see Chapter 10).
Sheridan and colleagues also made the observation that heparin concentrations strongly influenced platelet activation by HIT sera: therapeutic (0.05-1U/mL), but not high (10-100 U/mL), heparin concentrations resulted in platelet activation, i.e., the characteristic "two-point" serotonin release activation profile of HIT. Later, Greinacher and colleagues (1994) showed that high heparin concentrations in solution release platelet factor 4 (PF4) from PF4-heparin complexes bound covalently to a solid phase, with a corresponding decrease in binding of HIT antibodies to the surface. Thus, the inhibition of platelet activation by high heparin concentrations probably results from a similar disruption of the multimolecular antigen complex on the platelet surface.
The high sensitivity of washed platelets to activation by HIT antibodies led to new insights into the pathogenesis of platelet activation. For example, 2 yr after describing their washed platelet assay for HIT, Kelton and coworkers (1988) reported that the platelet activation process was critically dependent on the platelet Fc receptor. This represented a fundamental new pathobiological mechanism in a drug-induced thrombocytopenic disorder.
Although several prospective studies of the frequency of HIT were performed (see Chapter 3), until the 1990s, none had systematically evaluated serum or plasma from study participants for HIT antibodies. Often the distinction between "early" and "late" thrombocytopenia was blurred. Thus, the relative frequency and clinical importance of immune versus nonimmune HIT were unclear. This is illustrated by a prospective study reported by Powers and colleagues (1979) that found HIT to be "uncommon" during treatment with porcine mucosal heparin, as "only" 4 of 120 (3%) patients developed thrombocytopenia, in contrast with the 26-31% frequency of thrombocytopenia reported for bovine lung heparin. However, 2 of these 120 patients probably died as a result of HIT-associated thrombosis (Warkentin and Kelton, 1990), underscoring the need for a specific laboratory marker for this immune-mediated syndrome.
In a prospective study of HIT that performed systematic testing for HIT antibodies (Warkentin et al., 1995), the authors showed the dramatic clinical effects of HIT. Of 665 patients participating in a clinical trial of UFH versus low molecular weight heparin (LMWH) after orthopedic surgery, nine patients developed "late" thrombocytopenia serologically confirmed to represent HIT. These patients had a thrombotic event rate far greater than controls. Moreover, the spectrum of thrombosis in HIT patients included venous thromboembolism, rather than only the classic problem of arterial thrombosis. This study also showed that early postoperative thrombocytopenia occurred frequently, but was not explained by HIT antibodies (see Chapter 3).
However, even this study did not initially capture the complete clinical profile of HIT. This is because it defined the platelet count fall indicating possible HIT using the "standard" definition of thrombocytopenia, i.e., a platelet count fall to less than 150 X 109/L (Warkentin et al., 1995). Subsequent review of the database, together with correlative analysis of the results of systematic serological testing for HIT antibodies (performed in most study subjects), showed that this standard definition underestimated the number of patients who had HIT (Warkentin et al., 2003). Rather, a proportional fall in platelet count (50% or greater)—in relation to the postoperative peak platelet count—provided a more accurate definition of thrombocytopenia (applicable at least to this postoperative patient population). This improved definition identified twice as many patients as having had HIT in this clinical trial, without compromising diagnostic specificity. Indeed, the study suggested that the risk of immune HIT is about 5% (16/332 = 4.8%) in postoperative orthopedic surgery patients receiving UFH for a week or more (see Chapter 3).
VI. THE TARGET ANTIGEN OF HIT: PF4-HEPARIN
In 1992, Jean Amiral, working in the laboratory of Dominique Meyer, reported that the antigen recognized by HIT antibodies was a complex between heparin and PF4, an endogenous platelet a-granule protein (Amiral et al., 1992). This important discovery led to an explosion of basic studies in numerous laboratories that led to further characterization of the basic pathogenesis of HIT (see Chapters 4-9). Amiral's discovery also led to the development of new assays for HIT antibodies based on enzyme immunoassay techniques (see Chapter 10).
The antigen site(s) recognized by HIT antibodies were identified as being on PF4, rather than on heparin itself or a compound antigen (Li et al., 2002) (see Chapters 5-7). This observation highlights intriguing parallels between HIT and the antiphospholipid syndrome. This latter disorder is also characterized by pathogenic antibodies directed against one or more proteins that express neoepi-topes when bound to certain negatively charged phospholipid surfaces (see Chapter 11). The presence of neoepitopes on the "self" protein, PF4, suggests that HIT can be conceptualized as a transient, drug-induced, platelet- and coagulation-activating autoimmune disorder. Indeed, high-titer HIT antibodies that are able to activate platelets in vitro even in the absence of pharmacologic heparin have been associated with the onset of thrombocytopenia and thrombosis beginning several days after heparin has been discontinued, so-called delayed-onset HIT (Warkentin and Kelton, 2001) (see Chapter 2).
The treatment of HIT is discussed in Chapters 12-21. Here we will discuss only a few vignettes relating to the initial use of selected treatments for HIT.
In 1982, a 48-yr-old vacationing American developed deep venous thrombosis and pulmonary embolism following a transatlantic flight to Germany. Heparin treatment was complicated by thrombocytopenia and progression of venous thrombosis. Professor Job Harenberg of Heidelberg University, who had performed phase I evaluations of the experimental glycosaminoglycan anticoagulant danaparoid, requested this agent from the manufacturer (NV Organon, The Netherlands). The platelet count recovered and the venous thrombosis resolved (Harenberg et al., 1983, 1997). Over the next 6 yr, this patient developed recurrent thromboembolic events, and was successfully treated each time with danaparoid. This favorable experience led to a named-patient, compassionate-release program ending in March 1997, during which time, over 750 patients were treated with this agent. Additionally, Chong and colleagues (2001) performed the first randomized, controlled clinical trial evaluating danaparoid (see Chapter 13).
B. Recombinant Hirudin (Lepirudin)
The medicinal leech, Hirudo medicinalis, has been used for medical purposes for many centuries. Given the observation that the medicinal leech can prevent clotting of blood it has ingested, crude preparations derived from this animal were given experimentally at the beginning of the twentieth century. However, because this treatment's daily cost (75 Reichsmark) in 1908 was equivalent to the monthly salary of a factory worker, it was judged to be infeasible. After World War I, Haas, at Justus-Liebig University in Giessen, began his experiments using crude extracts of leech heads for hemodialysis. The major complication in these animal experiments was severe bleeding. The first human hemodialysis patients were treated by him with hirudin during dialysis when a more purified, but still crude protein extract of leech heads became available (Haas, 1925).
In 1956, Dr. F. Markwardt began his work to extract the active component of the leech at the Ernst-Moritz-Arndt University, in Greifswald. Still today, elderly peasants in the small villages around Greifswald tell stories of how they earned their pocket money by collecting leeches for the researchers at the nearby medical school.
The production of large amounts of hirudin by recombinant technology allowed assessment of this direct thrombin inhibitor in clinical trials. Dr. Andreas Greinacher, at that time working at the Justus-Liebig University in Giessen, first used a recombinant hirudin (lepirudin) to anticoagulate a patient who developed acute HIT following heart transplantation. After Greinacher's move to Greifswald, he further assessed the use of hirudin in patients with HIT in two clinical studies that led to the first approval of a drug for parenteral anticoagulation of patients with HIT in both the European Community (March 1997) and the United States (March 1998) (Greinacher et al., 1999) (Table 1).
A theme of this book is the central importance of increased thrombin generation in the pathogenesis of thrombosis complicating HIT. The recognition that warfarin therapy can be deleterious in some patients with HIT illustrates the importance of uncontrolled thrombin generation in HIT.
TABLE 1 U.S. Approvals for Three Direct Thrombin Inhibitors
HIT indications For patients with HIT and associated thromboembolic disease in order to prevent further thromboembolic complications For prophylaxis or treatment of thrombosis in patients with HIT Anticoagulation in patients with or at risk for HIT undergoing PCI For patients with or at risk of HIT/HITTS undergoing PCI
Non-HIT indications Use as an anticoagulant in patients with unstable angina undergoing PTCA Use (with provisional use of GP IIb/IIIa inhibitor) as an anticoagulant in patients undergoing PCI
Date of U.S. approval Lepirudin Argatroban Bivalirudin
June 30, 2000 April 3, 2002
November 30, 2005
December 15, 2000 June 13, 2005
March 6, 1998
Abbreviations: GP, glycoprotein; HIT, heparin-induced thrombocytopenia; HITTS, heparin-induced thrombocyto-penia/thrombosis syndrome; PCI, percutaneous coronary intervention; PTCA, percutaneous transluminal coronary angioplasty.
In December 1992, in Hamilton, Canada, while receiving ancrod and warfarin treatment for deep vein thrombosis complicating HIT, a 35-yr-old woman developed progressive venous ischemia, culminating in venous limb gangrene. This occurred despite a supratherapeutic international normalized ratio (INR). The following day, Kelton observed an area of skin necrosis on the abdomen of this patient, suggesting the diagnosis of warfarin-induced skin necrosis. The author questioned whether the warfarin had also contributed to the pathogenesis of the venous limb gangrene. This hypothesis was directly tested just 2 mo later when a second young woman developed severe phlegmasia cerulea dolens of an upper limb during treatment of deep vein thrombosis complicating HIT with ancrod and warfarin. Treatment with vitamin K and plasma given by pheresis reversed the phlegmasia. Further laboratory studies supported this hypothesis of a disturbance in procoagulant-anticoagulant balance during treatment of HIT with warfarin (Warkentin et al., 1997) (see Chapters 2,11, and 12).
Increasingly, HIT became viewed as a syndrome characterized by multiple prothrombotic events, including not only platelet and endothelial cell activation, but also profound activation of coagulation pathways. This conceptual framework provides a rationale for antithrombotic therapy that reduces thrombin generation in patients with HIT (Warkentin et al., 1998).
Was this article helpful?